1 23
Ecosystems
 
ISSN 1432-9840
 
Ecosystems
DOI 10.1007/s10021-014-9830-0
Soil Development and Nutrient Availability
Along a 2 Million-Year Coastal Dune
Chronosequence Under Species-Rich
Mediterranean Shrubland in Southwestern
Australia
Benjamin L. Turner & Etienne Laliberté
1 23
Your article is protected by copyright and all
rights are held exclusively by Springer Science
+Business Media New York. This e-offprint is
for personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.
Soil Development and Nutrient
Availability Along a 2 Million-Year
Coastal Dune Chronosequence Under
Species-Rich Mediterranean
Shrubland in Southwestern
Australia
Benjamin L. Turner1* and Etienne Laliberte´2
1Smithsonian Tropical Research Institute, Apartado 0843-03092, Balboa, Ancon, Republic of Panama; 2School of Plant Biology,
The University of Western Australia, Crawley, Perth, Western Australia 6009, Australia
ABSTRACT
Soil chronosequences provide valuable model
systems to investigate pedogenesis and associated
effects of nutrient availability on biological com-
munities. However, long-term chronosequences
occurring under seasonally dry climates remain
scarce. We assessed soil development and nutrient
dynamics along the Jurien Bay chronosequence, a
2 million-year sequence of coastal dunes in
southwestern Australia. The chronosequence is
significant because it occurs in a Mediterranean
climate and supports hyperdiverse shrublands
within a global biodiversity hotspot. Young soils
formed during the Holocene (<6,500 years old)
are strongly alkaline and contain abundant car-
bonate, which is leached from the profile within a
few thousand years. Middle Pleistocene soils (ca
120,000–500,000 years old) are yellow decalcified
sands with residual iron oxide coatings on quartz
grains over a petrocalcic horizon that occurs at
increasing depth as soils age. Early Pleistocene
soils (>2,000,000 years old) are completely lea-
ched of iron oxides and consist of bleached quartz
sand several meters deep. Changes in soil organic
matter and nutrient status along the Jurien Bay
chronosequence are consistent with patterns
observed along other long-term chronosequences
and correspond closely to expectations of the
Walker and Syers (1976) model of biogeochemical
change during pedogenesis. Organic carbon and
nitrogen (N) accumulate rapidly to maximum
amounts in intermediate-aged Holocene dunes
and then decline as soils age. In contrast, total
phosphorus (P) declines continuously along the
chronosequence to extremely low levels after
2 million years of pedogenesis, eventually repre-
senting some of the lowest P soils globally. Ratios
of soil organic carbon to P and N to P increase
continuously along the chronosequence, consis-
tent with a shift from N limitation on young soils
to extreme P limitation on old soils. Phosphorus
fractionation by sequential extraction reveals a
rapid decline in primary and non-occluded phos-
phate and an increase in organic and occluded P as
soils age. Concentrations of extractable (that is,
Received 25 June 2014; accepted 20 October 2014
Electronic supplementary material: The online version of this article
(doi:10.1007/s10021-014-9830-0) contains supplementary material,
which is available to authorized users.
Author contributions BT and EL designed the study and conducted
the fieldwork. BT conducted laboratory analysis and wrote the manu-
script, with input from EL.
*Corresponding author; e-mail: TurnerBL@si.edu
Ecosystems
DOI: 10.1007/s10021-014-9830-0
 2014 Springer Science+Business Media New York
Author's personal copy
readily bioavailable) N and P, as well as
exchangeable cations, are greatest in Holocene
dunes and decline to low levels in Pleistocene
dunes. Extractable micronutrient concentrations
were generally very low and varied little across
the chronosequence. We conclude that the Jurien
Bay chronosequence is an important example of
changing patterns of nutrient limitation linked to
long-term soil and ecosystem development under
a Mediterranean climate.
Key words: chronosequence; ecosystem devel-
opment; pedogenesis; nutrients; phosphorus;
nitrogen.
INTRODUCTION
A soil chronosequence is a series of soils that differ
from each other only in the time since the onset of
their formation (Stevens and Walker 1970; Vito-
usek 2004). Other soil-forming factors, including
parent material, topography, climate, and vegeta-
tion, are held relatively constant, although it is
recognized that climate and vegetation rarely con-
form strictly to this definition (Huggett 1998).
Nevertheless, soil chronosequences provide an
important means of assessing pedogenesis and
associated patterns of nutrient availability and
limitation in terrestrial ecosystems over timescales
from hundreds to millions of years (Vitousek 2004;
Walker and others 2010; Turner and Condron
2013).
Typically, nutrient availability follows a predict-
able pattern during long-term pedogenesis (Walker
and Syers 1976). Nitrogen (N) is absent from most
parent materials, but accumulates quickly in young
soils through biological N fixation (Menge and
Hedin 2009). In contrast, phosphorus (P) declines
continually as pedogenesis proceeds, as it is lost in
runoff at a greater rate than it is supplied by bed-
rock weathering and atmospheric deposition
(Walker and Syers 1976). At the same time,
chemical and biological transformations further
reduce the availability of the P remaining in the soil
by converting primary mineral phosphate into or-
ganic forms and occluding them within secondary
minerals (Walker and Syers 1976; Parfitt and others
2005; Turner and others 2007). As a result, primary
productivity tends to be limited by N availability on
young soils and by P availability on old soils
(Vitousek and Farrington 1997; Laliberte´ and oth-
ers 2012; Coomes and others 2013). In very old
soils on stable landscapes, P limitation can become
strong enough to constrain plant biomass, produc-
tivity, and other ecosystem processes, termed
‘ecosystem retrogression’ (Wardle and others
2004), and to influence the distribution and
diversity of plant and microbial communities
(Williamson and others 2005; Lambers and others
2008; Jangid and others 2013; Laliberte´ and others
2013a, 2014).
Long-term retrogressive soil chronosequences
have been identified in only a few locations around
the world (Peltzer and others 2010). Well-known
examples are the Hawaiian Islands chronosequence
(Crews and others 1995; Vitousek 2004), the Franz
Josef post-glacial chronosequence in New Zealand
(Walker and Syers 1976), and the Cooloola coastal
dune chronosequence in eastern Australia
(Thompson 1981). However, the majority of these
chronosequences occur in relatively humid cli-
mates, with only a small number in more arid
ecosystems. These include the 3 million-year San
Joaquin alluvial terrace sequence under Mediter-
ranean grassland in California (Harden 1982;
Baisden and others 2002) and the 3 million-year
basalt chronosequence in the Northern Arizona
Volcanic Field, which supports a semiarid pin˜on–
juniper woodland (Selmants and Hart 2008, 2010).
Long-term nutrient dynamics might be expected
to deviate from the Walker and Syers (1976) model
in dry ecosystems through a slower rate of P loss by
leaching (Lajtha and Schlesinger 1988). However,
both the Northern Arizona and the San Joaquin
chronosequences show patterns of soil nutrients
that are consistent with the Walker and Syers
model of nutrient dynamics during pedogenesis.
Additional information from similar arid environ-
ments around the world, which includes much of
the Australian continent, is required to confirm the
broad applicability of the Walker and Syers (1976)
model to Mediterranean ecosystems.
Here we report the results of a study of soils
along a 2 million-year coastal dune chronose-
quence at Jurien Bay, Western Australia. The
chronosequence occurs under a Mediterranean
climate but differs from the previously studied dry
chronosequences in terms of its origin and parent
material (coastal carbonate dunes) and is of par-
ticular interest because of the exceptionally high
levels of plant diversity due to its location within
a global biodiversity hotspot (the Southwest
B. L. Turner and E. Laliberte´
Author's personal copy
Australian Floristic Region) (Hopper and Gioia
2004). A glasshouse nutrient-addition experiment
using soils from the Jurien Bay chronosequence
found evidence for a shift from N to P limitation
(Laliberte´ and others 2012), which was supported
by a subsequent analysis of foliar nutrient con-
centrations and resorption efficiency in naturally
occurring plants in the field (Hayes and others
2014). Here we report the results of a detailed
assessment of (i) pedogenesis through soil profile
descriptions and nutrient budgets, (ii) the dynamics
of organic matter, (iii) amounts and composition of
soil P, and (iv) biologically available nutrient con-
centrations, including N, P, major cations, and
micronutrients.
THE JURIEN BAY CHRONOSEQUENCE
The 2 million-year Jurien Bay dune chronose-
quence is located in southwestern Australia,
approximately 200 km north of Perth (3001¢–
3024¢ south, 11457¢–11511¢ east; Figure 1A).
The chronosequence is part of the Swan Coastal
Plain, where a series of coastal sand deposits run
parallel to the coastline for approximately 400 km
from Busselton in the south to Jurien Bay in the
north (McArthur and Bettenay 1974; McArthur
2004). The sand deposits were formed by sea level
fluctuations since the Early Pleistocene (or possibly
Late Pliocene), approximately 2.59 million years
ago (Kendrick and others 1991; Laliberte´ and oth-
ers 2012). The Swan Coastal Plain lies within the
Perth Basin, which extends approximately
1000 km north–south along the coast of south-
western Australia. The Perth Basin is bounded to
the east by the Darling Fault, which separates it
from the Archean Yilgarn Block, part of the Wes-
tern Australian Shield (Playford and others 1976).
Coastal dune soils in the Swan Coastal Plain have
been grouped into three main units based on the
corresponding dune systems (McArthur and Bet-
tenay 1974; McArthur 2004; Laliberte´ and others
2012). The three dune systems (Quindalup,
Spearwood, and Bassendean) correspond generally
with the Safety Bay Sand, the Tamala Limestone,
and the Bassendean Sand, respectively (Table 1)
(Playford and others 1976).We delineated six main
chronosequence stages within these three dune
systems (see below). The Spearwood and Bassen-
dean dunes are more than 20 km wide in places
along the Swan Coastal Plain, although the Bas-
sendean dunes disappear abruptly north of the
Jurien Bay area. Below we describe the Jurien Bay
chronosequence in the context of the major soil-
forming factors of time, parent material, climate,
vegetation, and topography (Jenny 1941), as well
as the influence of disturbance on the integrity of
the chronosequence.
Ages of Dune Formation
Sea level changes during the Quaternary are now
well established based on marine oxygen isotope
records (for example, Lambeck and Chappell 2001;
Berger 2008; Lewis and others 2013) and provide
the basis for understanding coastal dune succession
along the Jurien Bay chronosequence. Sea level
has lowered progressively since the Miocene,
punctuated by periodic sea level high stands asso-
ciated with interglacial periods (Miller and others
Figure 1. Map of Australia (A), showing the location of the Jurien Bay chronosequence, and a detailed map of the Jurien
Bay area (B), showing the three main dune systems (Quindalup, Spearwood, and Bassendean) and the locations of the six
profile pits. Note that the Eatha, Nylagarda, and Yerramullah soils are not part of the chronosequence.
Soil Development and Nutrient Availability
Author's personal copy
2005). The major dune units that occur along the
Jurien Bay chronosequence most likely correspond
with the greatest sea level high stands, with dunes
formed during lower-level high-stands being
reworked or submerged by subsequent high stands
(Brooke and others 2014).
Coastal dunes in southwestern Australia represent
three main phases of dune development associated
with global sea level high stands during the Holocene,
Middle Pleistocene, and Early Pleistocene (Table 1).
Even older sand deposits, the Yoganup Formation,
occur in the region at the foot of the Darling Escarp-
ment (Kendrick and others 1991). The formation is
probably older than Pliocene but is not well under-
stood and is not included in the Jurien Bay chrono-
sequence, which extends only to deposits ranging in
age from the present to the Early Pleistocene, or
possibly Late Pliocene (Kendrick and others 1991).
The oldest deposits along the Jurien Bay chrono-
sequence are the Bassendean dunes. These have
been assigned an Early Pleistocene age based on their
interpretation as the dune facies of the underlying
Ascot Formation, which has been dated by marine
fossil assemblages (Kendrick and others 1991).
The Spearwood dunes formed during the Middle
Pleistocene, ca 120,000–500,000 years ago. Based
on heavy mineral assemblages, five Spearwood
dune sub-systems have been recognized in the
Perth area (with increasing age from west to east,
the Trigg, Karrinyup, Gwelup, Balcatta, and Yokine
Dunes) (Bastian 1996), although they have proven
difficult to date precisely (Hewgill and others 1983;
Murray-Wallace and Kimber 1989; Price and others
2001; Hearty and O’Leary 2008). However, there is
widespread agreement that the youngest Spear-
wood dune system corresponds to the last inter-
glacial sea level high stand (Marine Isotope Stage
(MIS) 5e, ca 116,000–128,000 years ago) (Stirling
and others 1998), and the increasing age of the
dune systems with progression inland is supported
by several lines of evidence, including heavy min-
eral assemblages and dating by thermolumines-
cence, optically stimulated luminescence, electron-
spin resonance, and amino-acid racemization
(Bastian 1996; Price and others 2001; Hearty and
O’Leary 2008; Brooke and others 2014). Older
dunes presumably correspond to earlier major sea
level high stands, which occurred at approximately
220,000 years (MIS 7), 330,000 years (MIS 9),
410,000 years (MIS 11), and 480,000 years (MIS
13) (Lisiecki and Raymo 2005). The upper age of
the Spearwood dunes is constrained at about
500,000 years by optical luminescence dating
(Brooke and others 2014). As sea level high stands
prior to about 500,000 years were relatively low, it
is presumed that older deposits (but younger than
the Bassendean dunes) were below the current sea
level and subsequently destroyed by Middle Pleis-
tocene high stands (Hearty and O’Leary 2008).
The Quindalup dunes (McArthur and Bettenay
1974) are the most recent dune system, formed in
the last 6,500 years since the Holocene post-glacial
sea level high stand (Woods and Searle 1983).
Although at least four phases of Quindalup dune
development have been identified in areas of the
Swan Coastal Plain (McArthur 2004), we grouped
the Holocene dunes at Jurien Bay into three age
classes based on the degree of soil development and
landscape position (Hayes and others 2014). Young
Quindalup dunes are no more than 100 years old,
have been only recently stabilized by vegetation,
and exhibit little or no pedogenic development.
Medium-aged Quindalup dunes are approximately
1,000 years old, have continuous vegetation cover,
and exhibit weak but recognizable development of
an epipedon in the soil profile. Old Quindalup
Table 1. The Three Main Phases of Dune Development Along the Jurien Bay Chronosequence, South-
western Australia
Dune
system
Dune
ages
Chronosequence
stages
Years
before
present
Marine
isotope
stage
Associated
geological
formation
Nature of soils
Quindalup Holocene 1, 2, 3 <6,500 1 Safety
Bay Sand
Calcareous sand
Spearwood Middle
Pleistocene
4, 5 120,000–
500,000
5e to 13 Tamala
Limestone
Decalcified sand
with residual iron
oxides over indurated
limestone
Bassendean Early
Pleistocene
(and possibly
Late Pliocene)
6 >2,000,000 – Ascot
Formation
Residual quartz sand;
parent material probably
contained less carbonate
than Quindalup
or Spearwood
B. L. Turner and E. Laliberte´
Author's personal copy
dunes are approximately 6,500 years old and occur
inland at the Holocene–Pleistocene transition. The
oldest dunes correspond to the final stages of the
post-glacial marine transgression (Lewis and others
2013).
Parent Material
The Holocene Quindalup dunes consist primarily of
carbonaceous material from the nearshore coastal
environment. The continental shelf off the coast of
the Swan Coastal Plain is extensive and shallow,
which promotes carbonate production and move-
ment onshore during marine transgressions
(Brooke and others 2014). The Spearwood dunes
are thought to have formed from weathering of the
underlying Tamala Limestone, a medium-to-coarse
grained aeolian calcarenite (calcareous sandstone)
composed of sand-sized shell fragments and quartz
grains, with minor amounts of feldspars and heavy
minerals including ilmenite, garnet, amphibole,
and epidote, a similar composition to minerals in
the Quindalup dunes (Bastian 1996). The feldspars
weather rapidly and disappear, as do garnet,
amphibole, and epidote, whereas ilmenite weath-
ers to leucoxene but does not disappear (Bastian
1996). Eventually, all unstable minerals disappear,
leaving almost pure quartz sand. Carbonate con-
tent in the Tamala Limestone varies from 30 to
96% (Bastian 1996; Hearty and O’Leary 2008;
Brooke and others 2014).
It has been suggested that Bassendean sands
might have originated as calcarenites in the same
way as the Spearwood and Quindalup dunes
(McArthur and Bettenay 1974; Playford and others
1976; Bastian 1996). However, it seems more likely
that the Bassendean Sand contained a lower car-
bonate content initially, based on its association
with the underlying Ascot formation (Kendrick and
others 1991). This might have been due either to
low carbonate production in the coastal environ-
ment at the time of deposition, or dilution by sili-
ceous sediments from the continental landmass
(Kendrick and others 1991). The extremely
weathered nature of the Bassendean Sand means
that it has been difficult to quantify its origin with
certainty, although reference soils from the three
dune systems contain similar particle size distribu-
tions, suggesting a common aeolian origin (McAr-
thur 2004). We do not consider the possibility of a
lower initial carbonate content in the Bassendean
Sand to compromise studies of plant–soil interac-
tions along the chronosequence (Laliberte´ and
others 2013b; Hayes and others 2014), because
carbonate is lost rapidly during the early stages of
pedogenesis in carbonate-rich soils (for example,
Salisbury 1925; Burges and Drover 1953; see
below) and the younger Spearwood dunes can
contain many meters of completely decalcified sand
over limestone. We therefore assume that any
carbonate initially present in the older Bassendean
dunes would have long disappeared from the soil
profile to great depth, so that the Bassendean dunes
represent the endpoint of an exceptionally strong
natural soil nutrient gradient for studies where
slight differences in parent material can be
neglected (Laliberte´ and others 2013b).
There have been persistent claims that the
Spearwood sands originate from inland dunes that
expanded onto the coastal plain during periods of
aridity (that is, the desert extension hypothesis;
Glassford and Killigrew 1976; Hearty and O’Leary
2008). However, this hypothesis has been refuted
on the basis of a number of lines of evidence
(reviewed in Laliberte´ and others 2013b), including
the complete absence of ‘desert sand’ in the vast
intervening areas between the purported desert
source and the Swan Coastal Plain destination
(Lowry 1977; Bastian 1996), a clear west-to-east
trend of increased weathering of the Spearwood
sands (Bastian 1996) that matches a west-to-east
increase in the age of the underlying limestone
(Price and others 2001; Hearty and O’Leary 2008;
Brooke and others 2014), and the match in particle
size distribution and mineralogy between the
Spearwood dune sand and sand embedded in the
underlying limestone (Wyrwoll and King 1984;
Tapsell and others 2003). These separate lines of
evidence confirm that the Spearwood sand is
residual and originates from in situ weathering and
decalcification of the Tamala Limestone, with some
local remobilization but no long-range transport.
Climate
The climate of the study area is Mediterranean,
with hot, dry summers and cool, moist winters.
Mean annual rainfall (measured between 1953 and
2012) is 573 mm, of which about 80% falls be-
tween May and September (Australian Bureau of
Meteorology, http://www.bom.gov.au/climate/
data/). There is little if any variation in rainfall
along the chronosequence, which extends only an
approximate 15 km inland from the coastline. The
mean annual temperature is 19C and the mean
annual maximum temperature is 25C, with Feb-
ruary being the warmest month (mean monthly
maximum of 31C) and July the coolest (mean
monthly maximum of 20C) (data from 1969 to
2013).
Soil Development and Nutrient Availability
Author's personal copy
In the USDA Soil Taxonomy system (Soil Survey
Staff 1999), the soil-moisture regime at Jurien Bay
qualifies as Xeric (that is, the soil is dry for ‡45 con-
secutive days in summer and moist for at least 45
consecutive days in winter), and the soil-tempera-
ture regime qualifies as Thermic (that is, the mean
annual soil temperature is 15–22C, and the differ-
ence between means of summer and winter tem-
peratures at 50 cm below the soil surface is ‡6C).
There is evidence that rainfall varied markedly in
inland areas of Western Australia during glacial
maxima, and included periods of increased aridity
(for example, Fujioka and others 2009). However,
the currently available evidence, although limited,
suggests that climate along the coastal margins of
southwestern Australia has been well buffered
from conditions inland, with relatively little varia-
tion from the modern climate over the time span
covered by the Jurien Bay dune chronosequence
(Wyrwoll and others 2014).
Vegetation
Vegetation in the Jurien Bay region is Mediterra-
nean low shrubland known as kwongan (Beard
1984; Hopper 2014). The chronosequence is in the
Southwest Australian Floristic Region (Hopper and
Gioia 2004), which is listed as a global biodiversity
hotspot (Myers and others 2000). Plant communi-
ties are dominated by schlerophyllous shrubs and
trees with remarkably high species richness and
endemism (Lamont and others 1984; Hopper and
Gioia 2004). There is a marked increase in plant
species diversity with soil age along the chronose-
quence (Laliberte´ and others 2013a, 2014). Fami-
lies such as Fabaceae and Myrtaceae are common
on younger dune systems, while Proteaceae
become more common on older dunes (Laliberte´
and others 2012; Hayes and others 2014).
In agreement with evidence for a relatively stable
climate along the coastal margins of southwestern
Australia, there is evidence that shrublands have
persisted in the region throughout the Last Glacial
Maximum, whereas changes in vegetation and
climate were more pronounced in the other regions
of Australia that now share a similar Mediterranean
climate (Dodson 2001; Barrows and Juggins 2005).
Some authors have suggested that the relative cli-
mate stability in southwestern Australia (at least
near the coastline) may explain why this region has
retained a hyperdiverse sclerophyll flora (Hopper
and Gioia 2004), whereas sclerophyll diversity was
greatly reduced in southeastern Australia due to
climate fluctuations during the Pleistocene
(Sniderman and others 2013).
Dune Morphology
Dune morphology varies along the chronose-
quence in terms of dune height and form, reflecting
differences in age of the three main dune systems.
The Quindalup dunes occur as marked ridges and
crests with short steep slopes. In particular, the
oldest Quindalup dunes form a prominent ridge
that marks the transition to the much older
Spearwood dunes (Laliberte´ and others 2012). The
Spearwood dunes have clear topographic succes-
sion (see above; Bastian 1996), but dune mor-
phology becomes subdued with distance inland.
The youngest Spearwood dunes marking the
Holocene–Pleistocene transition are the most
prominent, with broad flattened but distinct dune
ridges further inland. The Bassendean dunes retain
some of their dune expression around Perth (Bas-
tian 1996), but have generally lost their charac-
teristic dune morphology, forming extensive flatter
areas between the younger Spearwood dunes and
the older laterite gravel slopes (the Peron Slopes) to
the east. However, the apparently buffered climatic
conditions over the Quaternary (Wyrwoll and
others 2014) and the absence of ‘rejuvenated’
profiles in Bassendean sand (see below) suggest
that the older dunes have not experienced recent
extensive erosion or reworking other than at the
local scale.
Disturbance
Southwestern Australia has not been glaciated
since the Permian (that is, at least 250 million years
ago) (Geological Survey of Western Australia
1990). The Australian continent separated from
India in the Cretaceous and from Antarctica in the
Eocene, and has been drifting northward ever since
(Playford and others 1976). The region is very
stable geologically (Wyrwoll and others 2014),
although some authors have argued that local
tectonism along the coast of southwestern Australia
may have influenced dune stratigraphy (Semeniuk
and Searle 1986; Playford and others 2013). In
general, however, the dunes have remained rela-
tively undisturbed in their depositional settings
other than local surface reworking. Importantly,
they have not been buried by younger sediments
and have therefore been exposed to weathering
since their deposition.
Fire is the main disturbance to vegetation in the
area, with typical fire-return intervals of <30 years
(Griffin and Burbidge 1990). For example, a study
on the Eneabba Sandplain, around 40 km from the
Jurien Bay chronosequence, estimated a mean fire-
B. L. Turner and E. Laliberte´
Author's personal copy
return interval of 13 years in the last 30 years
(Enright and others 2005). In these fire-prone
shrublands, seedling recruitment occurs almost
exclusively after fire from seeds that are either
stored in the canopy (that is, serotiny; Lamont and
others 1991) or as a persistent soil seed bank (Bell
and others 1993). In addition, many plant species
in these shrublands can re-sprout vigorously after a
fire (Bell and others 1984).
We do not know of any evidence for a substantial
dust imprint on the sequence (Laliberte´ and others
2013b), which is perhaps unsurprising given the
prevailing westerly air flow from the Indian Ocean.
We cannot discount greater dust inputs during
glacial periods when low sea levels exposed the
continental shelf. However, the extremely low total
P concentrations on the oldest (Bassendean) soils,
including in surface soil layers (see below), suggest
the absence of long-term dust inputs. Marine aer-
osol inputs are presumably greater at the coast and
decrease inland, although this is not reflected in
exchangeable sodium concentrations, which are
generally very low and are not the greatest in the
youngest Quindalup soils closest to the coastline
(see below).
METHODS
Soil Profile Pits and Plot Selection
Profile pits were excavated in August 2012 at a
number of sites along the chronosequence and six of
these, most clearly representing the main chrono-
sequence stages, are described in detail here (Fig-
ure 1B; Table 1; see description of chronosequence
stages below). Profiles were excavated to at least 1 m
deep and up to 2 m deep for older, more developed
soils. Pits were located on the shoulders or upper
slopes of dunes, except for the Bassendean profile
where dune morphology was indistinct. Profiles
were described and sampled by genetic horizon.
Profile descriptions and analytical data are presented
in full in Supplementary Online Material.
To quantify changes in soil nutrients relevant to
potential limitation of biological activity, we also
sampled surface soils (0–20 cm depth) from 60
10 m 9 10 m plots across the chronosequence (ten
plots per chronosequence stage) that were selected
using a random-stratified sampling design (Lali-
berte´ and others 2014). Details on plot selection
and delineation of the chronosequence stages are
provided elsewhere (Laliberte´ and others 2012,
2014; Hayes and others 2014) and are summarized
here. Aerial photographs, maps (including soil,
geology, and topography), and on-ground verifi-
cation (for example, using soil morphological
characteristics) were used to define six broad re-
gions (chronosequence stages) corresponding to
distinct dune systems of increasing age, as indicated
by the degree of soil development (McArthur
2004). Stages 1–3 can be easily distinguished from
stages 4–6 because the latter occur east of a distinct
beach ridge that marks the transition between the
Holocene and Pleistocene dunes. In addition, soils
of stages 1–3 are still alkaline due to the presence of
carbonates, whereas soils from stages 4–6 are
decalcified (see below). Within the Holocene dunes
(stages 1–3), mobile dunes (stage 1) were easily
identified by their recent origin and absence of soil
development. Dunes corresponding to stage 2 were
located between dunes of stage 1 and 3, but are
heavily vegetated, and their soils are much more
developed (with an organically enriched A hori-
zon) than those of stage 1. Older Holocene dunes
(stage 3) were selected on the basis of landscape
position (that is, just west of the Holocene–Pleis-
tocene transition). Soils from stages 4–5 (Spear-
wood dunes) are easily distinguished by their
distinct yellow color that is due to the oxidation of
iron minerals. Stages 4 and 5 consist of Middle to
Late Pleistocene dunes with a calcarenite core
(Tamala Limestone); they show calcretes and karst-
weathering profiles in the upper parts that are
overlain by decalcified sand, the thickness of which
depends on soil age and the degree of weathering
(Bastian 1996). In stage 4, calcrete is found within
1 m of the soil surface, whereas in stage 5, it occurs
at a depth of several meters or is absent in some
cases (see below). Finally, dunes corresponding to
stage 6 (Bassendean dunes) can be distinguished
from stages 4–5 by their lack of morphological
expression, gray (not yellow) sands, thick (>2 m)
bleached eluvial horizons, and the absence of cal-
crete.
These six stages were spread over a 675 km2
(that is, approximately 45 km north–south 9
15 km east–west) coastal area from south of Jurien
Bay to Green Head. The regions were digitized as
polygons in a geographic information system
(Quantum GIS, http://qgis.org/). All vehicle-
accessible roads and tracks from the general area
were digitized into the GIS. Randomly positioned
points within each of the six distinct dune systems
were generated in the GIS, constraining locations
between 40 and 200 m from a road or track to
facilitate access. Due to the relatively small sizes of
vegetation patches within the youngest dune sys-
tem (that is, stage 1: young mobile dunes), loca-
tions were randomly selected without
consideration to road or track proximity. Ten plots
Soil Development and Nutrient Availability
Author's personal copy
per dune system were selected for vegetation sur-
veys, for a total of 60 plots. Plots were selected only
in areas that showed no evidence of recent fires
(that is, in the last 5 years). The average closest
distance between two neighboring plots was
2.1 km. In each plot, we identified an additional
seven randomly located 2 m 9 2-m subplots. Be-
tween June 11 and 16, 2012, seven soil samples
(one sample per 2-m 9 2-m subplot) were col-
lected in each of the 60 100 m2 plots, for a total of
420 samples. Subsampling was done to obtain more
accurate estimates of the mean. Samples were ta-
ken to 20-cm depth using a 5-cm diameter sand
auger, sieved (<2 mm) to remove roots and other
large organic debris, and then air dried prior to
laboratory analyses (see below).
Soil Analytical Procedures
Samples from soil profile pits were air dried and
analyzed for particle size distribution, pH, carbon-
ate, organic C, total N, total P, and exchangeable
cations. Samples from the 60 10 m 9 10 m plots
were analyzed for total C, total N, total P, sequen-
tial P fractionation, micronutrients, carbonate, pH,
exchangeable phosphate, extractable N, N miner-
alization, and exchangeable cations. Soil pH was
determined in both deionized water and 10 mM
CaCl2 in a 1:2 soil-to-solution ratio using a glass
electrode. The concentrations of sand (53 lm–
2 mm)-, silt (2 lm–53 lm)-, and clay (<2 lm)-
sized particles were determined by the pipette
method following pretreatment to remove soluble
salts and organic matter (Gee and Or 2002). Total C
and N were determined by automated combustion
and gas chromatography with thermal conductivity
detection using a Thermo Flash 1112 analyzer (CE
Elantech, Lakewood, New Jersey). Total P was
determined by ignition (550C, 1 h) and extraction
in 1 M H2SO4 (16 h, 1:50 soil to solution ratio),
with phosphate detection by automated neutral-
ization and molybdate colorimetry on a Lachat
Quikchem 8500 (Hach Ltd, Loveland, CO).
Exchangeable cations were determined by extrac-
tion in 0.1 M BaCl2 (2 h, 1:30 soil-to-solution ra-
tio), with detection by inductively coupled plasma
optical-emission spectrometry (ICP–OES) on an
Optima 7300 DV (Perkin-Elmer Ltd, Shelton,
Connecticut) (Hendershot and others 2008). Car-
bonate content was determined by mass loss after
addition of 3 M HCl (Loeppert and Suarez 1996),
and organic C was calculated as the difference be-
tween total C and carbonate-C. Bulk density was
determined by taking three replicate cores of
known volume per horizon using a 7.5-cm diam-
eter stainless steel ring and determining the soil
mass after drying at 105C.
Readily exchangeable phosphate was determined
by extraction with anion-exchange membranes
(AEM) (Turner and Romero 2009). Soil (5 g on a
dry-weight basis) was shaken for 24 h with 80 ml
deionized water and five anion-exchange resin
strips (1 9 4 cm; manufactured by BDH, Poole,
UK, and distributed by VWR International, West
Chester, Pennsylvania). The strips were rinsed in
deionized water and the phosphate recovered by
shaking for 1 h in 50 ml of 0.25 M H2SO4. Phos-
phate was determined in the acid solution at
880 nm following online neutralization with NaOH
by automated molybdate colorimetry using a La-
chat Quikchem 8500. Nitrogen was extracted from
fresh samples within 8 h of collection. Soil was
shaken in 1 M KCl and the solution was filtered
through a Whatman No. 42 filter paper. A simple
measure of N mineralization was determined by
incubation of field-moist soils for 7 days at ambient
laboratory temperature and extraction in 1 M KCl,
with changes in ammonium and nitrate deter-
mined by difference from the corresponding con-
centrations in fresh soils extracted on the day of
collection. Ammonium was determined by auto-
mated colorimetry at 660 nm following reaction
with phenolate, while nitrate was determined at
520 nm following cadmium-catalyzed reduction to
nitrite and reaction with sulfanilamide at pH 8.5.
Both ammonium and nitrate were determined
using a Lachat Quikchem 8500 flow injection
analyzer. Total dissolved N was determined by
combustion and gas chromatography on a Shima-
dzu TOC-TN analyzer, and dissolved organic N was
calculated as the difference between the total dis-
solved N and the sum of ammonium and nitrate.
Total exchangeable bases (TEB) was calculated as
the sum of Ca, K, Mg, and Na; effective cation-
exchange capacity (ECEC) was calculated as the
sum of Al, Ca, Fe, K, Mg, Mn, and Na; base satu-
ration was calculated by (TEB  ECEC) 9 100.
Micronutrients (Cu, Fe, Mn, and Zn) were ex-
tracted in Mehlich-III solution (Mehlich 1984) with
detection by ICP–OES.
Soil P was fractionated in samples from the 60
10 m 9 10 m plots using a modified sequential
extraction procedure (Hedley and others 1982).
Soils were extracted sequentially with AEM, 0.5 M
NaHCO3, 0.1 M NaOH, and 1 M HCl. Residual P
was determined on the soil after the four extrac-
tions using the ignition and acid-extraction proce-
dure described above. In all cases, extracts were
centrifuged (8,0009g, 15 min) and decanted for
analysis. Each aliquot was neutralized using phe-
B. L. Turner and E. Laliberte´
Author's personal copy
nolphthalein indicator and dilute NaOH or H2SO4
(as appropriate) and analyzed for inorganic phos-
phate, and total P. Phosphate was determined by
molybdate colorimetry at 880 nm with a 1-cm path
length. Total P was determined by the same pro-
cedure following acid-persulfate digestion at 80C
overnight in sealed glass tubes. In both cases,
standards were prepared in the extract solution
following identical neutralization and dilution
steps. The detection limit for both procedures was
approximately 0.6 mg P kg-1 except for resin P
(<0.1 mg P kg-1). Organic P was calculated as the
difference between total P and inorganic phos-
phate. Fractions were grouped into four pools: or-
ganic P (organic P extracted in NaHCO3 and
NaOH), primary mineral P (inorganic P extracted in
HCl), non-occluded P (inorganic P extracted in
NaHCO3 and NaOH), and occluded P (that is,
residual P).
RESULTS
Soil Development Along the Jurien Bay
Chronosequence
Examples of soil profiles corresponding to the six
stages of the Jurien Bay chronosequence are shown
in Figure 2, with key diagnostic features and pro-
file-weighted soil properties reported in Table 2.
Detailed profile descriptions and full chemical and
Figure 2. Soil development along the Jurien Bay chronosequence. Profiles represent the six main chronosequence stages
studied: A young Quindalup (<100 years old), B medium-aged Quindalup (1,000 years old), C old Quindalup
(6,500 years old); D young Spearwood (120,000 years old), E old Spearwood (500,000 years old), and F Bassendean
(>2,000,000 years old). Note the absence of horizon development in the youngest Quindalup profile (A), the marked
color change in the old Quindalup profile (C) as carbonates are removed, the yellow colors indicating residual iron oxides
over shallow limestone in the young Spearwood profile (D), the incipient eluvial horizon in the old Spearwood profile (E),
and the complete removal of iron oxides in the Bassendean profile (F).
Soil Development and Nutrient Availability
Author's personal copy
physical analyses by genetic horizon are provided
in the Online Supplementary Material.
Soils developed on Quindalup dunes (stages 1–3;
Figure 2A–C) are light gray-to-grayish brown and
exhibit weakly developed surface horizons en-
riched with organic matter overlying several meters
of unweathered calcareous sand. The Quindalup
dunes lack diagnostic horizons and are classified as
Typic Xeropsamments, reflecting the Xeric mois-
ture regime and sand texture. In the Australian Soil
Classification System (Isbell 2002), these soils are
Shelly Rudusols. They consist of calcarenites (that
is, sand-sized carbonate grains), with up to 82%
carbonate in the youngest profile (Table 2). The
two youngest dunes qualify as carbonatic (>40%
carbonate), but the old Quindalup dunes show
evidence of decalcification, with about 25% car-
bonate throughout the profile. Although much
more developed than the younger profiles, the old
Quindalup dune does not qualify as an Inceptisol,
because the B horizons does not meet all criteria for
a cambic horizon. Specifically, although there is
evidence of decalcification and a higher color value
than underlying horizons, the texture was slightly
coarser than very fine sand. All the Quindalup
profiles were strongly or very strongly alkaline
throughout.
Carbonate in the young Quindalup dunes
weathers rapidly and leaches from the profile.
Thus, the youngest Spearwood profile (stage 4;
Figure 2D) exhibited complete decalcification in
the upper horizons, leaving brownish yellow sub-
soil consisting of greater than 95% residual quartz
sand and iron oxide coatings. The yellow color
derives from the presence of goethite and weath-
ering of heavy minerals such as ilmenite, garnet,
and amphibole (Bastian 1996). The young Spear-
wood profile (stage 4) contained an undulating
indurated petrocalcic horizon (that is, a calcrete cap
over the underlying Tamala Limestone), reaching
between 40 and 100 cm below the soil surface. The
presence of the petrocalcic horizon qualifies the
profile as an Inceptisol (Calcic Haploxerepts), con-
sistent with the much older and pedogenically
developed nature of the profile. The soil is very
friable and has very weak blocky structure.
The depth of sand and depth to limestone in
Spearwood dunes increases from west to east,
consistent with an increasing degree of weathering
and leaching of carbonate. However, we did not
encounter limestone in the older Spearwood profile
(stage 5; Figure 2E). As a result, the older Spear-
wood profile classifies as an Entisol (Xeric Quart-
zipsamment). In the Australian Soil Classification
system, the Spearwood profiles are Yellow-Orthic
Tenosols, with the younger profile qualifying as a
Petrocalcic Great Group. The greater age of the
older Spearwood profile was indicated by an
incipient eluvial horizon, which becomes much
more pronounced in Spearwood profiles further
inland (and therefore, presumably older) and at
wetter sites further south (Turner and Laliberte´,
unpublished data). This bleached horizon indicates
the early stages of iron oxide removal from the
profile as iron–humus complexes and demonstrates
that these older Spearwood soils are transitioning
toward the final stage of soil development along the
chronosequence (stage 6; Figure 2F). Soil pH in
water in the Spearwood profiles was neutral and
ECEC was less than 2 cmolc kg
-1. The young and
old Spearwood soils along the Jurien Bay chrono-
sequence are broadly equivalent to the yellow and
gray phases, respectively, of the Karrakatta sands
(McArthur and Bettenay 1974).
The final and oldest stage of the chronosequence
(stage 6) is represented by soils developed on the
Bassendean dunes (Figure 2). The Bassendean
profile is at least 2 million years old (that is, Early
Pleistocene) and represents the ultimate end point
of pedogenesis along the chronosequence. A thin
organic horizon occurs at the surface, consisting of
partly decomposed litter and cluster roots from the
common nearby trees in the Proteaceae (for
example, Banksia attenuata, Banksia menziesii). The
mineral soil is composed entirely of white-to-light
gray single grain quartz sand. The sand is com-
pletely free of iron oxides, which have been lea-
ched from the soil as iron–organic matter
complexes. These complexes are often deposited as
a spodic horizon (that is, a subsurface accumulation
of humus–metal complexes and metal oxides) in
the deep subsoil (McArthur 2004), forming a
brown iron–humus material known locally as
‘coffee rock.’ Such soils qualify as Giant Humus
Podzols in the Australian Soil Classification
(Thompson 1992), but in many cases (for example,
Prakongkep and others 2012), they do not qualify
as Spodosols in Soil Taxonomy because the spodic
horizon is more than 200 cm below the soil surface.
At Jurien Bay, a transition to very pale brown
sand, but not a spodic horizon, was encountered at
more than 300 cm (further augering was prevented
by the water table), although on well-drained areas
elsewhere we have encountered spodic materials at
several meters depth on Bassendean sands. The
nature of the spodic horizon varies markedly
depending on proximity to the water table
(McArthur and Russell 1978). Despite its great age,
the absence of diagnostic features means that the
Bassendean profile classifies as an Entisol (Xeric
B. L. Turner and E. Laliberte´
Author's personal copy
T
ab
le
2.
S
o
il
P
ro
fi
le
s,
A
p
p
ro
x
im
at
e
A
ge
s,
T
ax
o
n
o
m
ic
C
la
ss
es
,a
n
d
P
ro
fi
le
-w
ei
gh
te
d
S
o
il
P
h
ys
ic
al
an
d
C
h
em
ic
al
P
ro
p
er
ti
es
fo
r
S
ix
K
ey
S
ta
ge
s
A
lo
n
g
th
e
Ju
ri
en
B
ay
C
h
ro
n
o
se
q
u
en
ce
,
S
o
u
th
w
es
te
rn
A
u
st
ra
li
a
C
h
ro
n
o
se
q
u
en
ce
st
ag
e
A
p
p
ro
x
im
at
e
ag
e
S
o
il
ta
x
o
n
o
m
y
P
ed
o
m
o
rp
h
ic
fe
at
u
re
s
C
ar
b
o
n
at
e
(%
)
S
an
d
(%
)
S
il
t
(%
)
C
la
y
(%
)
p
H
(H
2
O
)
p
H
(C
aC
l 2
)
E
C
E
C
(c
m
o
l c
k
g-
1
)
1
Q
u
in
d
al
u
p
yo
u
n
g
1
0
0
C
ar
bo
n
at
ic
,
th
er
m
ic
,
T
yp
ic
X
er
o
p
sa
m
m
en
ts
In
ci
p
ie
n
t
A
h
o
ri
zo
n
8
2
9
8
.2
1
.1
0
.7
9
.1
8
.2
1
2
.9
3
2
Q
u
in
d
al
u
p
m
ed
iu
m
1
0
0
0
C
ar
bo
n
at
ic
,
th
er
m
ic
,
T
yp
ic
X
er
o
p
sa
m
m
en
ts
D
ev
el
o
p
m
en
t
o
f
A
h
o
ri
zo
n
6
6
9
5
.9
2
.1
2
.0
9
.2
8
.1
5
.5
2
3
Q
u
in
d
al
u
p
o
ld
6
5
0
0
S
il
ic
eo
u
s,
th
er
m
ic
,
T
yp
ic
X
er
o
p
sa
m
m
en
ts
M
o
d
er
at
e
d
ec
al
ci
fi
ca
ti
o
n
,
d
ev
el
o
p
m
en
t
o
f
w
ea
k
st
ru
ct
u
re
2
5
9
6
.9
1
.4
1
.6
8
.9
8
.0
5
.0
1
4
S
p
ea
rw
o
o
d
yo
u
n
g
1
2
0
,0
0
0
(M
IS
5
e)
S
an
d
y,
si
li
ce
o
u
s,
th
er
m
ic
,
C
al
ci
c
H
ap
lo
x
er
ep
ts
D
ec
al
ci
fi
ca
ti
o
n
an
d
d
ep
o
si
ti
o
n
as
an
in
d
u
ra
te
d
p
et
ro
ca
lc
ic
h
o
ri
zo
n
;
re
si
d
u
al
ir
o
n
o
x
i-
d
es
th
ro
u
gh
o
u
t
th
e
p
ro
fi
le
0
9
4
.0
2
.9
3
.1
6
.9
6
.1
1
.9
4
5
S
p
ea
rw
o
o
d
o
ld
4
8
0
,0
0
0
(M
IS
1
3
?)
T
h
er
m
ic
,
X
er
ic
Q
u
ar
tz
ip
sa
m
m
en
ts
D
ec
al
ci
fi
ca
ti
o
n
;
re
si
d
u
al
ir
o
n
o
x
id
es
;
in
ci
p
ie
n
t
el
u
vi
al
h
o
ri
zo
n
;p
et
ro
ca
lc
ic
h
o
ri
zo
n
at
co
n
si
d
er
ab
le
d
ep
th
(i
f
p
re
se
n
t
at
al
l)
0
9
6
.7
1
.3
2
.0
6
.8
5
.7
0
.9
6
6
B
as
se
n
d
ea
n
>
2
,0
0
0
,0
0
0
T
h
er
m
ic
,
X
er
ic
Q
u
ar
tz
ip
sa
m
m
en
ts
C
o
m
p
le
te
re
m
o
va
l
o
f
ir
o
n
o
x
id
es
,
le
av
in
g
re
si
d
u
al
q
u
ar
tz
sa
n
d
0
9
7
.3
2
.3
0
.3
5
.6
4
.4
0
.9
8
M
IS
=
m
ar
in
e
is
ot
op
e
st
ag
e;
E
CE
C
=
ef
fe
ct
iv
e
ca
ti
on
ex
ch
an
ge
ca
pa
ci
ty
.
Soil Development and Nutrient Availability
Author's personal copy
Quartzipsamments) in Soil Taxonomy. In the
Australian Soil Classification system, the Bassen-
dean profile studied here is a Grey-Orthic Tenosol.
The Bassendean soils are equivalent to the Janda-
kot sands (McArthur and Bettenay 1974), although
they appear to lack the spodic horizon around Ju-
rien Bay. Soil pH in water was moderately acid, and
ECEC was less than 1 cmolc kg
-1.
Profile-Weighted Organic Carbon, Total
Nitrogen, and Total Phosphorus
Organic C in the upper meter of soil increased
throughout the Holocene dune sequence (stages
1–3), reaching a maximum of 17.7 kg C m-2 in the
oldest Quindalup profile (stage 3) (Table 3;
Figure 3A). Organic C stocks then declined in
Table 3. Profile-Weighted Nutrient Concentrations for the Upper Meter of Soil in Profile Pits of the Six Main
Stages Along the Jurien Bay Chronosequence, Southwestern Australia
Chronosequence stage Organic C
(g m-2)
Total N
(g m-2)
Total P
(g m-2)
C:N C:P N:P
1 Quindalup young 8,900 495.5 384.3 18.0 23.2 1.3
2 Quindalup medium 14,678 763.3 346.4 19.2 42.4 2.2
3 Quindalup old 17,720 362.2 194.8 48.9 90.9 1.9
4 Spearwood young 3,864 166.5 28.8 23.2 134.4 5.8
5 Spearwood old 2,917 119.2 12.7 24.5 228.9 9.4
6 Bassendean 4,063 117.8 6.4 34.5 639.2 18.5
Figure 3. Profile-
weighted contents of
organic carbon (A), total
nitrogen (B), and total
phosphorus (C) to one-
meter depth along the
Jurien Bay
chronosequence,
southwestern Australia.
B. L. Turner and E. Laliberte´
Author's personal copy
the Spearwood profiles (2.9–3.9 kg C m-2), but
increased again to 4.1 kg C m-2 in the oldest pro-
file (stage 6). Total N (approximating total organic
N) followed a slightly different pattern, peaking at
763 g N m-2 in the medium-aged Quindalup dune
(stage 2), before declining to less than 120 g N m-2
in the oldest dunes (Table 3; Figure 3B).
Total P declined continuously along the chron-
osequence. In the Holocene dunes, total P declined
by 50% in about 6500 years, from a maximum of
384 g P m-2 in the youngest Quindalup profile
(stage 1) to 195 g P m-2 in the oldest Quindalup
profile (stage 3) (Table 3; Figure 3C). Total P then
declined markedly to 28.8 g P m-2 in the youngest
Spearwood profile and eventually to only
6.4 g P m-2 after 2 million years of pedogenesis
(stage 6).
Profile-weighted element ratios showed clear
patterns expected during long-term pedogenesis
(Table 3). The organic C to total N ratio increased in
Holocene dunes to a maximum of 48.9 in the old
Quindalup profile (stage 3), and then declined in
Spearwood dunes, although the ratio increased to
34.5 in the oldest profile (stage 6). The organic C to
total P ratio increased continuously throughout the
sequence, from a low of 23.2 in the youngest
Holocene dune (stage 1) to a maximum of 639 in
the oldest profile on Bassendean Sand (stage 6).
Similarly, the profile-weighted N:P increased from
2.2 or less in the Holocene dunes to 18.5 in the
oldest profile (stage 6).
Changes in pH and Carbonate in Surface
Soils
Soil pH in water declined from very strongly alkaline
(pH 9.2) in the youngest dunes to strongly alkaline
(pH 8.6) in older Quindalup dunes, and then to
moderately acid (pH 5.9) in the Bassendean soil
(Figure 4A). Soil pH in 10 mM CaCl2 followed a
similar pattern but approximately one pH unit low-
er, with values decreasing from 8.2 in the youngest
dunes to 4.8 in the oldest dunes. Carbonate in sur-
face soils was 76–78% in the youngest two dunes,
declining to 43% in the older Quindalup dunes, al-
though values for old dunes varied markedly among
replicate plots (Figure 4B). No carbonate was de-
tected in Spearwood or Bassendean dunes.
Changes in Carbon, Nitrogen, and
Phosphorus in Surface Soils
Organic C concentrations in surface soils increased
initially to a maximum of 18.8 g C kg-1 in the
medium-aged Quindalup dunes (stage 2) and then
declined to less than 7 g C kg-1 on Spearwood and
Bassendean dunes (stages 4–6; Figure 5A). Total N
followed a similar pattern, increasing to
1.17 g N kg-1 in the medium-aged Quindalup
dunes and declining to 0.30 g N kg-1 or less in the
older stages (Figure 5B). Total P increased initially
from 350 mg P kg-1 on the youngest dunes (stage 1)
to 433 mg P kg-1 on the medium-aged Quindalup
Figure 4. Changes in soil pH measured in deionized water (A) and carbonate (as CaCO3 equivalent) (B) in surface soils
along the Jurien Bay chronosequence, southwestern Australia. Values are from ten replicate plots per chronosequence
stage. The central vertical bar in each box shows the median, the box represents the interquartile range, the whiskers show the
location of the most extreme data points that are still within a factor of 1.5 of the upper or lower quartiles, and the black
points are values that fall outside the whiskers.
Soil Development and Nutrient Availability
Author's personal copy
dunes (stage 2; Figure 5C), reflecting redistribution
of P in surface horizons by plants (Figure 3). It then
declined to 21.5 mg P kg-1 on the Spearwood
dunes and then to 6.7 mg P kg-1 on the Bassen-
dean dunes (stage 6; Figure 5C).
The organic C to total N ratio in surface soil de-
clined initially from 27.0 in the youngest soils to
about 15.5 in older Quindalup soils, and then in-
creased to a maximum of 31.6 in old Spearwood
soils (Figure 5D). Organic C to total P ratios in-
creased markedly from very low values in
Quindalup soils ( £ 51 on stages 1–3) to around
1,000 in the oldest dunes (Figure 5E). The total N:P
ratio followed a similar pattern to the organic C:P
ratio, increasing consistently from very low values
on young soils (<4 on young Quindalup dunes) to
maximum values (41.6) on the oldest soils (Bas-
sendean dunes; stage 6) (Figure 5F).
Phosphorus Fractionation in Surface
Mineral Soils
Phosphorus fractions determined by sequential
extraction closely followed the classic Walker and
Syers (1976) model of P transformations during
pedogenesis (Figure 6). The youngest soils (stages
1–3) were dominated by HCl-extractable phosphate,
which occurred at up to 355 mg P kg-1 (Figure 6A)
and accounted for up to 90% of the total P (Fig-
ure 6B). The HCl-extractable phosphate pool de-
clined to low concentrations (<1 mg P kg-1) in
Spearwood and Bassendean soils (stages 4–6). The
majority of the remainder of the total P in the
Quindalup dunes was present as non-occluded
inorganic phosphate, which occurred at up to
35.7 mg P kg-1 and accounted for 4–8% of the total
P. Non-occluded inorganic phosphate (that is,
bicarbonate plus NaOH-extractable inorganic
phosphate) declined to low concentrations
(<1 mg P kg-1) in older soils, but still represented
11–23% of the total P. Occluded P (that is, residual P)
occurred at relatively minor concentrations
throughout the sequence (up to 29.2 mg P kg-1 in
the oldest Quindalup dunes; stage 3), but accounted
for a much greater proportion of the total P in older
soils (29–50%) than on younger soils (2–17%). The
decline in the proportion of occluded P in the oldest
soils appears to be related to the increasing relative
Figure 5. Concentrations of organic carbon (A), total nitrogen (B), total phosphorus (C), and their ratios (D–F) in surface
soils along the Jurien Bay chronosequence, southwestern Australia. Values are from ten replicate plots per chronosequence
stage. See legend to Figure 4 for an explanation of the box plots.
B. L. Turner and E. Laliberte´
Author's personal copy
importance of organic P under conditions of ex-
tremely low soil P. Organic P was only 2 mg P kg-1
in the youngest soil, increased to the maximum
concentration in the oldest Quindalup soil
(18.2 mg P kg-1; stage 3), and then declined to low
concentrations in Spearwood and Bassendean soils
(<3 mg P kg-1; stages 4–6). However, organic P
increased continuously as a proportion of the total P
throughout the sequence, reaching 40% in the
Bassendean soil (stage 6). Total P calculated from the
sum of sequential fractions corresponded closely
with total P values determined independently by
ignition and acid extraction (Figure 5C). The rela-
tionship when forced through the origin was de-
scribed by the equation Y = 1.03x (R2 = 0.98,
P < 0.0001), where Y = total P determined by
ignition and x = total P by sequential fractionation.
Extractable Phosphorus and Nitrogen in
Surface Mineral Soils
Phosphate extracted by AEM was greatest in the
medium-aged Quindalup dunes (2.1 mg P kg-1;
stage 2) and then declined to very low values on
the two oldest stages ( £ 0.5 mg P kg-1) (Fig-
ure 7A). Extractable N followed a similar pattern
(Figure 7B). Total dissolved N increased to maxi-
mum values on the medium-aged Quindalup
dunes (stage 2) and then declined to lower values
on older soils. Dissolved organic N accounted for
the majority of the total dissolved N, being
approximately 20 mg N kg-1 in medium and older
Quindalup dunes and no more than 10 mg N kg-1
elsewhere. Both extractable ammonium and ni-
trate increased from the lowest values in the young
Quindalup dunes (stage 1) to maximum values in
the medium-aged Quindalup dunes (stage 2), and
then declined to moderate values on older soils
(Figure 7B). Ammonium concentrations declined
following soil incubation, with losses mirroring
concentrations of extractable ammonium (that is,
greatest in the medium-aged Quindalup dunes).
Nitrate mineralization was greatest in medium and
old Quindalup dunes (0.3–0.5 mg N kg-1 day-1),
with negligible mineralization on the youngest and
oldest dunes (Figure 7C).
Exchangeable Cations in Surface Soils
Extractable Al concentrations were no more than
0.01 cmolc kg
-1 along the entire sequence, so
ECEC was dominated throughout by base cations
(Figure 7D), with base saturation around 100%.
Effective cation exchange capacity was the greatest
in the youngest Quindalup soils (24.9 cmolc kg
-1;
stage 1), declining by about 50% in older
Quindalup dunes (stages 2–3) and then further to
very low values in the older soils (a minimum va-
lue of 2.0 cmolc kg
-1 in old Spearwood soils; stage
5) (Figure 7D).
The exchange capacity was dominated by cal-
cium throughout the sequence, with maximum
values on the youngest Quindalup dunes
(23.0 cmolc kg
-1) and minimum values on the
oldest two stages (<2.0 cmolc kg
-1; stages 5–6).
Exchangeable magnesium declined from 1.81
cmolc kg
-1 in the youngest soil to 0.33 cmolc kg
-1
in old Spearwood soils (Figure 7D). Exchangeable
potassium concentrations were much lower,
with very low values (0.01 cmolc kg-1) on the
Figure 6. Phosphorus fractionation by sequential extraction in surface soils along the Jurien Bay chronosequence,
southwestern Australia, including concentrations (A) and proportions of the total phosphorus (B). Phosphorus is grouped
into primary mineral phosphate (extracted in HCl), organic phosphorus (organic phosphorus extracted in NaHCO3 and
NaOH), non-occluded phosphorus (inorganic phosphate extracted in NaHCO3 and NaOH), and occluded phosphorus (that
is, residual phosphorus not extracted in any step). Values are the means of ten replicate plots per chronosequence stage.
Soil Development and Nutrient Availability
Author's personal copy
youngest soils, increasing to about 0.05 cmolc kg
-1
on older Quindalup and young Spearwood soils
(stages 3–4), and then declining to low values (0.02
cmolc kg
-1) in the two oldest soils (stages 5–6;
Figure 7E). Exchangeable sodium followed a simi-
lar pattern, with maximum values in the medium-
aged Quindalup dunes ( 0.15 cmolc kg-1) and
declining to low values (0.03 cmolc kg
-1) on Bas-
sendean dunes (Figure 7F).
Micronutrient Concentrations in Surface
Mineral Soils
Of the four micronutrients measured, copper con-
centrations were extremely low throughout the
sequence relative to the amounts considered ade-
quate for crops, while manganese and zinc were
moderately low, and iron moderate to adequate
(for example, Lindsay and Norvell 1978; Adriano
1986) (Figure 8). Mehlich-III extractable copper
concentrations were the lowest in the youngest
Quindalup dunes (stage 1; 0.011 mg Cu kg-1) and
greatest in the youngest Spearwood dunes (stage 4;
0.043 mg Cu kg-1). Copper concentrations were
lowest, however, on the old Spearwood dunes
(stage 5; 0.008 mg Cu kg-1). Mehlich-III extract-
able iron concentrations were low in the Holocene
dunes (stages 1–3; <3 mg Fe kg-1) but much
greater in the older dunes (25–39 mg Fe kg-1).
Mehlich-III extractable manganese concentrations
followed a similar pattern to copper, increasing
from the lowest values on the youngest Quindalup
dunes (mean 0.71 mg Mn kg-1) to the greatest
concentrations on young Spearwood dunes (stage
4; 3.01 mg Mn kg-1). Mehlich-III extractable zinc
concentrations were greatest in the youngest
Quindalup dunes (stage 1; 1.43 mg Zn kg-1) but
changed little along the sequence. The lowest zinc
concentrations were detected in the Bassendean
dunes (stage 6; 0.54 mg Zn kg-1).
Figure 7. Extractable (readily bioavailable) nutrient concentrations in surface soils along the Jurien Bay chronosequence,
southwestern Australia, showing anion-exchange membrane extractable phosphate (A), KCl-extractable nitrogen (B),
mineralized nitrogen (C), effective cation-exchange capacity (ECEC) and the contribution of calcium, magnesium, and
other cations (D), exchangeable potassium (E), and exchangeable sodium (F). Values are the mean (±1 standard devia-
tion) of ten replicate plots per chronosequence stage. See legend to Figure 4 for an explanation of the box plots.
B. L. Turner and E. Laliberte´
Author's personal copy
DISCUSSION
Pedogenesis Along the Jurien Bay
Chronosequence
Soil development along the Jurien Bay chronose-
quence is consistent with previous detailed studies
in the Perth region that have shown a clear west-
to-east increase in weathering and soil develop-
ment (Bastian 1996; McArthur 2004). Soils with
ages between the oldest Quindalup and youngest
Spearwood dunes, a span of many thousands of
years, are not found along the sequence because
they have been buried by younger Holocene sedi-
ments or submerged beneath present sea level
(Brooke and others 2014). However, the Quinda-
lup and Spearwood successions, coupled with the
oldest Bassendean soil, are sufficient to reconstruct
the transition from calcareous sands to siliceous
quartz sands over 2 million years of pedogenesis.
Young soils are strongly alkaline, but undergo
relatively rapid acidification and leaching of car-
bonate. The carbonate precipitates at depth to form
a calcrete cap overlying the Tamala Limestone,
leaving an overlying layer of residual quartz sand.
With further pedogenesis, the calcite is progres-
sively leached away and residual iron oxides are
leached from the profile, leaving bleached quartz
grains. Depending on landscape position, the iron
oxides can form a spodic horizon consisting of iron–
humus complexes that may be many meters below
the soil surface. Although the leaching of iron
oxides is not pronounced along the Jurien Bay
chronosequence, an incipient eluvial horizon was
evident in profile pits on the most easterly Spear-
wood dunes, and we have studied eluvial horizons
up to 2-m deep in old Spearwood dunes under
greater rainfall further south.
Patterns of Soil Nutrients Along the
Chronosequence
Despite the modest rainfall under a Mediterra-
nean climate at Jurien Bay, changes in C, N, and
P along the chronosequence were remarkably
consistent with expectations of the Walker and
Syers (1976) model of nutrient transformations
during pedogenesis. Profile weights of organic C
and N increased rapidly in the early stages of
pedogenesis and then declined. In contrast, total
P declined continuously throughout the chrono-
sequence, reaching extremely low levels on the
oldest dunes. Indeed, total P concentrations in
Bassendean soils are among the lowest reported
globally. In parallel, profile-weighted C:P and N:P
ratios increased continuously along the sequence,
reinforcing the expectation of strengthening P
limitation in old soils (see below) (Laliberte´ and
others 2012; Hayes and others 2014). Changes in
soil P chemistry determined by sequential frac-
tionation were also in accordance with the
Walker and Syers (1976) model, with a marked
decline in primary mineral P during the early
stages of pedogenesis and increasing importance
of organic and occluded P in old soils. The
exception was the decline in the proportion of
occluded P in the oldest soils, which appears to
be due to the increasing importance of organic P
in those soils. This pattern of P transformations
during pedogenesis has been observed consis-
tently in chronosequences around the world un-
der a variety of climates and in soils developed
on markedly different parent materials (Walker
and Syers 1976; Crews and others 1995; Peltzer
and others 2010; Selmants and Hart 2010; Turner
and others 2012; Izquierdo and others 2013).
Figure 8. Concentrations of extractable micronutrients in surface soils along the Jurien Bay chronosequence, south-
western Australia, including copper (A), iron (B), manganese (C), and zinc (D). Values are from ten replicate plots per
chronosequence stage. See legend to Figure 4 for an explanation of the box plots.
Soil Development and Nutrient Availability
Author's personal copy
To assess changes in soil fertility along the se-
quence over a much broader area than possible
through soil profile pits, we determined concen-
trations of total and available nutrients in 100 m2
plots distributed along the six key chronosequence
stages. These samples further supported expecta-
tions of changes in nutrient availability and limi-
tation along the chronosequence (Laliberte´ and
others 2012; Hayes and others 2014). For example,
measures of N availability increased initially and
then declined, as did some cations (that is, potas-
sium and sodium) and micronutrients, whereas
extractable P, exchangeable calcium and magne-
sium, and extractable zinc declined continuously
along the sequence. The latter is somewhat sur-
prising given evidence for zinc limitation on the
youngest soils (Laliberte´ and others 2012), includ-
ing the high resorption efficiency for zinc in leaves
of native plants (Hayes and others 2014). This
result suggests that high pH, rather than low zinc
concentrations, is the cause of apparent zinc limi-
tation due to the slow diffusion of zinc in calcare-
ous soils (Melton and others 1973).
Copper deficiency is widespread on strongly
weathered soils of Western Australia, including on
acid and calcareous sands (Holloway and others
2008), and the extremely low copper concentra-
tions throughout the Jurien Bay chronosequence
are reflected in foliar copper concentrations far
below those considered adequate for crop growth
(Hayes and others 2014). Copper availability
therefore represents a potentially important
micronutrient constraint to plant performance
throughout the Jurien Bay chronosequence, al-
though we did not observe strong responses to
micronutrients on most soils in a previous phy-
tometer experiment (Laliberte´ and others 2012), or
high copper resorption in the leaves of native
plants (Hayes and others 2014). However, the
limited resorption might be explained by the fact
that plants with low foliar copper concentrations
tend to show low copper resorption efficiency from
leaves, because a greater proportion of the copper is
firmly bound in structural constituents such as cell
walls (White 2012).
Comparison with Other Long-Term
Chronosequences
The Jurien Bay chronosequence differs from
previously studied long-term chronosequences in a
number of important ways, including its carbonate
parent material, Mediterranean climate, and spe-
cies-rich shrubland vegetation. In terms of climate,
there are only two long-term chronosequences
under dry climates with which to compare the
Jurien Bay chronosequence. In the Northern Ari-
zona Volcanic Field, a 3 million-year chronose-
quence on basalt under 340-mm annual rainfall
and 11C mean annual temperature, total P in
surface soil declined by 60% during the lifetime of
the sequence (Selmants and Hart 2010). Primary
mineral P in surface soils declined by 75%, but was
still abundant (35% of the total P) in the oldest
soils. However, non-occluded P declined to low
concentrations and organic P increased continu-
ously throughout the sequence, although it only
accounted for 8% of the total P on the oldest soils.
In contrast, primary mineral P in apatite declined
almost completely after ca 100,000 years of pedo-
genesis after a 3 million-year chronosequence of
alluvial terraces supporting Californian grassland
under 300 mm annual rainfall and 16C mean
annual temperature (Baisden and others 2002). We
interpret the differences in the depletion of primary
mineral P between these two western USA se-
quences, despite similar rainfall, as an indication of
the importance of parent material in determining
long-term rates of P transformations, although this
does not account for possible differences in paleo-
climate. Although total P and other P forms were
not reported for the San Joaquin terrace sequence,
it appears to most closely resemble rates of P
depletion along the Jurien Bay chronosequence,
where primary mineral phosphate was completely
leached from the soil within 120,000 years. How-
ever, all three of these relatively dry chronose-
quences appear to conform to the Walker and Syers
(1976) model of P transformations during pedo-
genesis, although the rates of P transformations are
considerably slower than under more humid cli-
mates.
The Jurien Bay chronosequence is comparable in
some ways to the long-term coastal dune chrono-
sequence at Cooloola in eastern Australia
(Thompson 1981; Walker and others 1981). Al-
though the dates of dune formation at Cooloola are
poorly constrained, the chronosequence spans
approximately 700,000 years of pedogenesis, and it
seems likely that at least some of the dune systems
correspond in age to those at Jurien Bay. For
example, thermal luminescence dating gives an age
of 120,000 years for dune stage 5 at Cooloola (Te-
jan-Kella and others 1990). The Cooloola sequence
differs from Jurien Bay because it is in a humid
subtropical climate (1,500 mm of rainfall per year,
mostly falling in the summer) and the parent
material contains little carbonate (Thompson
1981). As a consequence, initial total soil P con-
centrations are much lower at Cooloola than at
B. L. Turner and E. Laliberte´
Author's personal copy
Jurien Bay: young vegetated dunes at Cooloola
contain less than 30 g P m-2 (Thompson 1981; B.L.
Turner, unpublished data), equivalent to the Jurien
Bay chronosequence after 120,000 years of pedo-
genesis (Table 3). However, despite the absence of
carbonate, soil development at Cooloola is quali-
tatively similar to that at Jurien Bay, with leaching
of iron oxide coatings on quartz grains leaving
bleached quartz profiles at least 2 m deep after
approximately 200,000 years of pedogenesis.
Eventually, after 400,000 years, the spodic horizon
occurs up to 20 m below the soil surface above a
profile of clean quartz sand (Thompson 1981,
1992), forming Giant Humus Podzols. This process
therefore occurs more rapidly than at Jurien Bay,
which is consistent with the greater productivity
(with tall eucalypt forest) and rainfall falling in the
warmest part of the year at Cooloola.
Most coastal dune chronosequences involve the
development of podzols, in which P depletion oc-
curs in parallel with the leaching of iron and alu-
minum from the upper part of the soil profile.
Stevens and Walker (1970) and Jenny (1941)
summarized a series of early studies of sand dune
chronosequences from Sweden, England, and else-
where. These and more recent studies of young
coastal dune sequences provide support for the
importance of rainfall in terms of podzol develop-
ment and P depletion (Syers and others 1970;
Lichter 1998; Tackett and Craft 2010). This is dem-
onstrated clearly by comparison of the Jurien Bay
chronosequence with Holocene coastal dune sys-
tems under a perhumid climate along the west coast
of the South Island of New Zealand (Wells and Goff
2007). These sequences began to form approxi-
mately 6500 years before present and therefore
span the same time period as the Quindalup dune
system at Jurien Bay. However, dune formation on
the west coast of New Zealand is triggered by
earthquakes along the Alpine Fault, which causes
landslides and sediment transport to the ocean and
subsequent dune formation. A well-studied exam-
ple occurs under temperate rainforest at Haast
(Turner and others 2012). Although total P in
young dunes at Haast is similar to young dunes at
Jurien Bay (300–400 mg P kg-1), rainfall is much
higher at Haast (3,500 mm year-1), so the rate of P
depletion is correspondingly greater. Comparing
concentrations in the upper mineral soil, total P
increased initially at Jurien Bay (through concen-
tration of P in topsoil by plants) and then declined to
around 200 mg P kg-1 after 6500 years. At Haast,
total P declined rapidly from about 320 mg P kg-1
initially to about 200 mg P kg-1 in just a few hun-
dred years, and then to about 100 mg P kg-1 after
about 4000 years. At that point, the majority of the
total soil P at Haast is in organic forms (>80%)
compared to about 10% at Jurien Bay, in agreement
with expectations for the influence of rainfall on soil
P composition (Walker and Syers 1976; Vitousek
2004). However, differences in vegetation are also
likely to be important, because the conifers that are
abundant in the early stages of the Haast chrono-
sequence produce leaf litter that decomposes slowly,
acidifies the soil rapidly, and promotes the forma-
tion of thick organic horizons within a few hundred
years (Coomes and Bellingham 2011; Turner and
others 2012).
The rate of decline of soil P at Haast is similar to
that at the nearby Franz Josef post-glacial chrono-
sequence, when the loss rate is expressed as a
proportion of the total P (Turner and others 2012).
Although these two sequences differ in their parent
material, they have formed under similar vegeta-
tion and climate. It is therefore surprising that the
proportional rate of total P decline at Franz Josef is
similar to that at Jurien Bay, despite large differ-
ences in parent material and climate (the Franz
Josef sequence is developed in graywacke and
schist deposits under 3,500–6,000-mm annual
rainfall). At Jurien Bay, total profile P to 1 m de-
clined from 384 to 28.8 g P m-2 in 120,000 years
(that is, the young Spearwood dunes), leaving
7.5% of the original P remaining. At Franz Josef,
total P in the upper 75 cm, including stones, de-
clined from 1,292 g m-2 initially to 76 g P m-2
after 120,000 years, leaving 5.9% of the total P
remaining (Walker and Syers 1976). This provides
further support for the suggestion that a consistent
pattern of proportional loss of P occurs among
chronosequences (Turner and others 2012).
Fire and Soil Nutrient Dynamics During
Ecosystem Development
Fire is an important driver of ecosystem nutrient
status, with implications for the rate of nutrient
transformations during pedogenesis (Peltzer and
others 2010). Fire causes the loss of N in gaseous
forms by volatilization from vegetation, litter, and
surface soil, but mineralizes organic P in live and
dead vegetation and returns it to the soil in inor-
ganic form (Carter and Darwin 2004; Certini 2005).
These processes act against the long-term pedo-
genic trend of increasing N and declining P, and
therefore effectively set the ecosystem back to an
earlier state in terms of soil nutrient stocks. It might
therefore be expected that systems in which fire is a
major and frequent disturbance, such as the kwon-
gan of southwestern Australia, would experience a
Soil Development and Nutrient Availability
Author's personal copy
slower rate of ecosystem development than less
fire-prone systems, perhaps even to the extent that
the long-term steady state P concentration is
greater in more fire-prone systems than in com-
parable areas that rarely experience fire. For
example, fires on Swedish Islands are of sufficient
intensity to burn peat, which maintains P avail-
ability and prevents the onset of retrogression on
the larger islands that burn frequently, compared to
small islands that almost never experience fire
(Wardle and others 2004). Fires are a common
disturbance along the Jurien Bay chronosequence,
but clearly this has not prevented the ecosystem
progressing toward extremely low P availability,
even on relatively young (that is, 120,000 year)
soils. This might be because fire promotes P loss in
this ecosystem, either as wind-blown ash, or when
rainfall onto burnt areas causes the released P (and
N) to be transported downslope in surface runoff or
in leachate through these sandy soils (Carter and
Darwin 2004; Certini 2005).
Ecological Consequences of Pedogenic
Change Along the Jurien Bay
Chronosequence
Pedogenesis and the associated changes in nutrient
status described here have important implications
for plant and microbial communities along the
Jurien Bay chronosequence. In particular, the re-
sults presented here provide further support for
experimental evidence suggesting that potential
nutrient limitation of plant growth along the
chronosequence corresponds closely to expecta-
tions of the Walker and Syers (1976) model, with N
limitation on young soils (perhaps along with other
nutrients such as potassium), co-limitation by
multiple nutrients on intermediate aged soils, and
strong P limitation on old soils (Laliberte´ and others
2012). This pattern is supported by evidence from
foliar nutrient analysis, with remarkably strong
community-wide declines in foliar P from young to
old soils, with concentrations spanning virtually
the global range of foliar P (0.1–3.0 mg P g-1 at
Jurien Bay) (Hayes and others 2014). In addition,
there is a marked increase in foliar N:P ratios along
the sequence (from 8 on young soils to 42 on oldest
soils) and a parallel increase in foliar P resorption
efficiency from zero to 79%, with all species con-
verging on an extremely conservative P use strat-
egy on the oldest soils (Hayes and others 2014).
Some Proteaceae, which become particularly
abundant on the oldest soils, express a number of
adaptations to acquire and conserve P, including
the formation of cluster roots (Lambers and others
2006), extremely high P use efficiency (Denton and
others 2007), low levels of protein and ribosomal
RNA (Sulpice and others 2014), and the replace-
ment of phospholipids by sulfolipids and galactoli-
pids during leaf development (Lambers and others
2012). Of particular interest is that plant diversity
increases continuously along the chronosequence
as soils age (Laliberte´ and others 2013a, 2014). This
conforms to a pattern observed previously in other
long-term soil chronosequences (Wardle and oth-
ers 2008), and primarily reflects the fact that plant
species pools (that is, sets of species from the re-
gional flora that can successfully grow in a given
habitat) on older acidic and P-impoverished soils
along the chronosequence are much larger than
those from younger alkaline soils (Laliberte´ and
others 2014). The species-rich flora in the Jurien
Bay region therefore provides an important
opportunity to study the influence of nutrient
availability and limitation on plant diversity and
distribution in a Mediterranean climate (Laliberte´
and others 2013a, 2014).
CONCLUSIONS
Soil development along coastal dune systems under
a Mediterranean climate in southwestern Australia
involves acidification and leaching of carbonate,
accumulation of organic matter, and a continual
decline in total P. Patterns of nutrient concentra-
tions correspond remarkably well with the Walker
and Syers (1976) model of nutrient transformations
and limitation during ecosystem development,
with a number of lines of evidence pointing to N
limitation on young soils (with possible micronu-
trient limitation), co-limitation by N and P on the
intermediate-aged soils, and strong P limitation on
old soils. Indeed, total P concentrations reach ex-
tremely low levels in the oldest soils and plant
communities demonstrate abundant signs of
adaptation to P stress (Hayes and others 2014). The
Jurien Bay chronosequence therefore represents an
important example of long-term pedogenesis and
associated changes in nutrient availability and
limitation in a Mediterranean climate. The species-
rich nature of the regional plant community offers
the possibility of linking the long-term conse-
quences of soil development and associated shifts in
nutrient availability and stoichiometry to patterns
of plant species distribution and diversity.
ONLINE SUPPLEMENTARY MATERIAL
Two online supplementary files contain (1)
descriptions of profiles corresponding to the six
B. L. Turner and E. Laliberte´
Author's personal copy
stages along the Jurien Bay chronosequence, and
(2) a table of physical and chemical data on soil
profiles by genetic horizon, including particle size
analysis, total carbon, nitrogen, and phosphorus,
carbonate, pH, and exchangeable cations. Data on
surface soils in the sixty 10 m 9 10 m plots are
available on request.
ACKNOWLEDGMENTS
Funding was provided by a Discovery Early Career
Researcher Award (DE120100352) from the Aus-
tralian Research Council and a Research Collabo-
ration Award from The University of Western
Australia (both to EL). The authors thank Dayana
Agudo, Pedro Arau´z, Aleksandra Bielnicka, and
Paola Escobar for laboratory support; and Felipe
Albornoz, Patrick Hayes, Hans Lambers, Kenny Png,
Franc¸ois Teste, Karl-Heinz Wyrwoll, and Graham
Zemunik for assistance in the field. Figure 1 was
created using base maps provided by the Depart-
ment of Agriculture and Food of Western Australia.
REFERENCES
Adriano DC. 1986. Trace elements in the terrestrial environ-
ment. New York: Springer.
Baisden WT, Amundson R, Brenner DL, Cook AC, Kendall C,
Harden J.2002. A multi-isotopeC and N modelinganalysisof soil
organic matter turnover and transport as a function of soil depth
in a California annual grassland soil chronosequence. Glob
Biogeochem Cycles 16:1135. doi:10.1029/2001GB001823.
Barrows TT, Juggins S. 2005. Sea-surface temperatures around
the Australian margin and Indian Ocean during the Last
Glacial Maximum. Quat Sci Rev 24:1017–47.
Bastian LV. 1996. Residual soil mineralogy and dune subdivi-
sion, Swan Coastal Plain, Western Australia. Aust J Earth Sci
43:31–44.
Beard JS. 1984. Biogeography of the kwongan. In: Pate JS,
Beard JS, Eds. Kwongan: plant life of the sandplain. Nedlands:
University of Western Australia Press. p 1–26.
Bell DT, Hopkins AJM, Pate JS. 1984. Fire in the kwongan. In:
Pate JS, Beard JS, Eds. Kwongan: plant life of the sandplain.
Nedlands: University of Western Australia Press. p 178–204.
Bell DT, Plummer JA, Taylor SK. 1993. Seed germination ecol-
ogy in southwestern Western Australia. Bot Rev 59:24–73.
Berger WH. 2008. Sea level in the late Quaternary: patterns of
variation and implications. Int J Earth Sci 97:1143–50.
Brooke BP, Olley JM, Pietsch T, Playford PE, Haines PW, Mur-
ray-Wallace CV, Woodroffe CD. 2014. Chronology of Qua-
ternary coastal aeolianite deposition and the drowned
shorelines of southwestern Western Australia—a reappraisal.
Quat Sci Rev 93:106–24.
Burges A, Drover D. 1953. The rate of podzol development in
sands of the Woy Woy District, NSW. Aust J Bot 1:83–94.
Carter MC, Darwin Foster C. 2004. Prescribed burning and
productivity in southern pine forests: a review. For Ecol
Manag 191:93–109.
Certini G. 2005. Effects of fire on properties of forest soils: a
review. Oecologia 143:1–10.
Coomes DA, Bellingham PJ. 2011. Temperate and tropical
podocarps: how ecologically alike are they? In: Turner BL,
Cernusak LA, Eds. Ecology of the Podocarpaceae in tropical
forests. Smithsonian Contributions to Botany. Washington,
DC: Smithsonian Institution Scholarly Press. p 119–40.
Coomes D, Bentley W, Tanentzap A, Burrows L. 2013. Soil
drainage and phosphorus depletion contribute to retrogressive
succession along a New Zealand chronosequence. Plant Soil
367:77–91.
Crews TE, Kitayama K, Fownes JH, Riley RH, Herbert DA,
Mueller-Dombois D, Vitousek PM. 1995. Changes in soil
phosphorus fractions and ecosystem dynamics across a long
chronosequence in Hawaii. Ecology 76:1407–24.
Denton M, Veneklaas EJ, Freimoser F, Lambers H. 2007. Banksia
species (Proteaceae) from severely phosphorus-impoverished
soils exhibit extreme efficiency in the use and re-mobilization
of phosphorus. Plant Cell Environ 30:1557–65.
Dodson JR. 2001. Holocene vegetation change in the mediter-
ranean-type climate regions of Australia. Holocene 11:673–
80.
Enright NJ, Lamont BB, Miller BP. 2005. Anomalies in grasstree
fire history reconstructions for south-western Australian
vegetation. Austral Ecol 30:668–73.
Fujioka T, Chappell J, Fifield LK, Rhodes EJ. 2009. Australian
desert dune fields initiated with Pliocene-Pleistocene global
climatic shift. Geology 37:51–4.
Gee GW, Or D. 2002. Particle size analysis. In: Dane JH, Topp C,
Eds. Methods of soil analysis, Part 4—Physical methods.
Madison: Soil Science Society of America. p 255–93.
Geological Survey of Western Australia. 1990. Geology and
Mineral Resources of Western Australia. Perth: Western
Australian Geological Survey. p 827.
Glassford DK, Killigrew LP. 1976. Evidence for Quaternary
westward extension of the Australian desert into south-wes-
tern Australia. Search 7:394–6.
Griffin EA, Burbidge AA. 1990. Description of the region. In:
Burbidge AA, Hopper DS, van Leeuwen S, Eds. Nature con-
servation, landscape and recreation values of the Lesueur
area. Perth: Environmental Protection Authority. p 15–24.
Harden JW. 1982. A quantitative index of soil development from
field descriptions: example from a chronosequence in central
California. Geoderma 28:1–28.
Hayes P, Turner BL, Lambers H, Laliberte´ E. 2014. Foliar nutri-
ent concentrations and resorption efficiency in plants of
contrasting nutrient-acquisition strategies along a 2-million-
year dune chronosequence. J Ecol 102:396–410.
Hearty PJ, O’Leary MJ. 2008. Carbonate eolianites, quartz sands,
and Quaternary sea-level cycles, Western Australia: a chro-
nostratigraphic approach. Quat Geochronol 3:26–55.
Hedley MJ, Stewart JWB, Chauhan BS. 1982. Changes in
inorganic and organic soil phosphorus fractions induced by
cultivation practices and by laboratory incubations. Soil Sci
Soc Am J 46:970–6.
Hendershot WH, Lalande H, Duquette M. 2008. Chapter 18. Ion
exchange and exchangeable cations. In: Carter MR, Gregorich
E, Eds. Soil sampling and methods of analysis. Boca Raton:
Canadian Society of Soil Science and CRC Press. p 173–8.
Hewgill FR, Kendrick GW, Webb RJ, Wyrwoll KH. 1983. Rou-
tine ESR dating of emergent Pleistocene marine units in
Western Australia. Search 14:215–17.
Soil Development and Nutrient Availability
Author's personal copy
Holloway RE, Graham RD, Stacey SP. 2008. Micronutrient deficien-
cies in Australian field crops. In: Alloway BJ, Ed. Micronutirent
deficiencies in global crop production. Berlin: Springer. p 63–92.
Hopper SD. 2014. Sandplain and kwongan: historical spellings,
meanings, synonyms, geography and definition. In: Lambers
H, Ed. Plant life on the sandplains in Southwest Australia, a
global biodiversity hotspot. Crawley: University of Western
Australia Publishing.
Hopper SD, Gioia P. 2004. The Southwest Australian Floristic
Region: evolution and conservation of a global hot spot of
biodiversity. Annu Rev Ecol Evol Syst 35:623–50.
Huggett RJ. 1998. Soil chronosequences, soil development, and
soil evolution: a critical review. Catena 32:155–72.
Isbell RF. 2002. The Australian soil classification. revised edn.
Collingwood: CSIRO Publishing.
Izquierdo JE, Houlton BZ, van Huysen TL. 2013. Evidence for
progressive phosphorus limitation over long-term ecosystem
development: examination of a biochemical paradigm. Plant
Soil 367:135–47.
Jangid K, Whitman WB, Condron LM, Turner BL, Williams MA.
2013. Soil bacterial community succession during long-term
ecosystem development. Mol Ecol 22:3415–24.
Jenny H. 1941. Factors of soil formation: a system of quantitative
pedology. New York: McGraw-Hill. p 281.
Kendrick GW, Wyrwoll K-H, Szabo BJ. 1991. Pliocene-Pleisto-
cene coastal events and history along the western margin of
Australia. Quat Sci Rev 10:419–39.
Lajtha K, Schlesinger WH. 1988. The biogeochemistry of phos-
phorus cycling and phosphorus availability along a desert soil
chronosequence. Ecology 69:24–39.
Laliberte´ E, Turner BL, Costes T, Pearse SJ, Wyrwoll K-H,
Zemunik G, Lambers H. 2012. Experimental assessment of
nutrient limitation along a 2-million-year dune chronose-
quence in the south-western Australia biodiversity hotspot. J
Ecol 100:631–42.
Laliberte´ E, Grace JB, Huston MA, Lambers H, Teste FP, Turner
BL, Wardle DA. 2013a. How does pedogenesis drive plant
diversity? Trends Ecol Evol 28:331–40.
Laliberte´ E, Turner BL, Zemunik G, Wyrwoll K-H, Pearse SJ,
Lambers H. 2013b. Nutrient limitation along the Jurien Bay
dune chronosequence: response to Uren & Parsons (2013). J
Ecol 101:1088–92.
Laliberte´ E, Zemunik G, Turner BL. 2014. Environmental fil-
tering explains variation in plant diversity along resource
gradients. Science 345:1602–5.
Lambeck K, Chappell J. 2001. Sea level change through the last
glacial cycle. Science 292:679–86.
Lambers H, Shane MW, Cramer MD, Pearse SJ, Veneklaas EJ.
2006. Root structure and functioning for efficient acquisition
of phosphorus: matching morphological and physiological
traits. Ann Bot 98:693–713.
Lambers H, Raven JA, Shaver GR, Smith SE. 2008. Plant
nutrient-acquisition strategies change with soil age. Trends
Ecol Evol 23:95–103.
Lambers H, Cawthray GR, Giavalisco P, Kuo J, Laliberte E, Pearse
SJ, Scheible W-R, Stitt M, Teste F, Turner BL. 2012. Proteaceae
from severely phosphorus-impoverished soils extensively re-
place phospholipids with galactolipids and sulfolipids during
leaf development to achieve a high photosynthetic phosphorus-
use-efficiency. New Phytol 196:1098–108.
Lamont BB, Hopkins AJM, Hnatiuk RJ. 1984. The flora—com-
position, diversity and origins. In: Pate JS, Beard JS, Eds.
Kwongan: plant life of the sandplain. Nedlands: University of
Western Australia Press. p 27–50.
Lamont BB, Maitre DCL, Cowling RM, Enright NJ. 1991. Can-
opy seed storage in woody plants. Bot Rev 57:277–317.
Lewis SE, Sloss CR, Murray-Wallace CV, Woodroffe CD, Smi-
thers SG. 2013. Post-glacial sea-level changes around the
Australian margin: a review. Quat Sci Rev 74:115–38.
Lichter J. 1998. Rates of weathering and chemical depletion in
soils across a chronosequence of Lake Michigan sand dunes.
Geoderma 85:255–82.
Lindsay WL, Norvell WA. 1978. Development of a DTPA soil test
for zinc, iron, manganese, and copper. Soil Sci Soc Am J
42:421–8.
Lisiecki LE, Raymo ME. 2005. A Pliocene-Pleistocene stack of 57
globally distributed benthic d18O records. Paleoceanography
20:PA1003.
Loeppert RH, Suarez DL. 1996. Carbonate and gypsum.
In: Sparks D et al., Eds. Methods in soil analysis, Part
3—Chemical methods. Madison: Soil Science Society of
America. p 437–74.
Lowry DC. 1977. Perth Basin yellow sand. Search 8:54–6.
McArthur WM. 2004. Reference soils of south-western Austra-
lia. Perth: Department of Agriculture. p 285.
McArthur WM, Bettenay E. 1974. Development and distribution
of soils of the Swan Coastal Plain, Western Australia. Can-
berra: CSIRO.
McArthur WM, Russell WGR. 1978. Soil morphological prop-
erties in relation to depth to groundwater table in a sandy
landscape near Perth. Aust J Soil Res 16:347–9.
Mehlich A. 1984. Mehlich 3 soil test extractant: a modification of
Mehlich 2 extractant. Commun Soil Sci Plant Anal 15:1409–16.
Melton JR, Mahtab SK, Swoboda AR. 1973. Diffusion of zinc in
soils as a function of applied zinc, phosphorus, and soil pH.
Soil Sci Soc Am J 37:379–81.
Menge DNL, Hedin LO. 2009. Nitrogen fixation in different
biogeochemical niches along as 120 000-year chronosequence
in New Zealand. Ecology 90:2190–201.
Miller KG, Kominz MA, Browning JV, Wright JD, Mountain GS,
Katz ME, Sugarman PJ, Cramer BS, Christie-Blick N, Pekar
SF. 2005. The Phanerozoic record of global sea-level change.
Science 310:1293–8.
Murray-Wallace CV, Kimber RWL. 1989. Quaternary marine
aminostratigraphy: Perth Basin, Western Australia. Aust J
Earth Sci 36:553–68.
Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB,
Kent J. 2000. Biodiversity hotspots for conservation priorities.
Nature 403:853–8.
Parfitt RL, Ross DJ, Coomes DA, Richardson SJ, Smale MC,
Dahlgren RA. 2005. N and P in New Zealand soil chronose-
quences and relationships with foliar N and P. Biogeochem-
istry 75:305–28.
Peltzer DA, Wardle DA, Allison VJ, Baisden WT, Bardgett RD,
Chadwick OA, Condron LM, Parfitt RL, Porder S, Richardson SJ,
Turner BL, Vitousek PM, Walker J, Walker LR. 2010. Under-
standing ecosystem retrogression. Ecol Monogr 80:509–29.
Playford PE, Cockbain AE, Lowe GH. 1976. Geology of the Perth
Basin, Western Australia. Bulletin 124 of the Geological Survey of
Western Australia. Perth: Geological Survey of Western Australia.
Playford PE, Cockbain AE, Berry PF, Roberts AP, Haines PW,
Brooke B. 2013. The geology of Shark Bay. East Perth: Geo-
logical Survey of Western Australia.
B. L. Turner and E. Laliberte´
Author's personal copy
Prakongkep N, Gilkes RJ, Singh B, Wong S. 2012. Pyrite and
other sulphur minerals in giant aquic spodosols, Western
Australia. Geoderma 181–182:78–90.
Price DM, Brooke BP, Woodroffe CD. 2001. Thermolumines-
cence dating of aceolianites from Lord Howe Island and south-
west Western Australia. Quat Sci Rev 20:841–6.
Salisbury EJ. 1925. Note on the edaphic succession in some dune
soils with special reference to the time factor. J Ecol 13:322–8.
Selmants PC, Hart SC. 2008. Substrate age and tree islands
influence carbon and nitrogen dynamics across a retrogressive
semiarid chronosequence. Glob Biogeochem Cycles. doi:10.
1029/2007GB003062.
Selmants PC, Hart SC. 2010. Phosphorus and soil development:
does the Walker and Syers model apply to semiarid ecosys-
tems? Ecology 91:474–84.
Semeniuk V, Searle DJ. 1986. Variability of Holocene sealevel
history along the southwestern coast of Australia—evidence
for the effect of significant local tectonism. Mar Geol 72:
47–58.
Sniderman JMK, Jordan GJ, Cowling RM. 2013. Fossil evidence
for a hyperdiverse sclerophyll flora under a non-Mediterra-
nean-type climate. Proc Natl Acad Sci 110:3423–8.
Soil Survey Staff. 1999. Soil taxonomy: A basic system of soil
classification for making and interpreting soil surveys. Lin-
coln: United States Department of Agriculture–Natural Re-
sources Conservation Service.
Stevens PR, Walker TW. 1970. The chronosequence concept and
soil formation. Q Rev Biol 45:333–50.
Stirling CH, Esat TM, Lambeck K, McCulloch MT. 1998. Timing
and duration of the Last Interglacial: evidence for a restricted
interval of widespread coral reef growth. Earth Planet Sci Lett
160:745–62.
Sulpice R, Ishihara H, Schlereth A, Cawthray GR, Encke B,
Giavalisco P, Ivakov A, Arrivault S, Jost R, Krohn N, Kuo J,
LalibertE´ E, Pearse SJ, Raven JA, Scheible W-R, Teste F,
Veneklaas EJ, Stitt M, Lambers H. 2014. Low levels of ribo-
somal RNA partly account for the very high photosynthetic
phosphorus-use efficiency of Proteaceae species. Plant Cell
Environ 37:1276–98.
Syers JK, Adams JA, Walker TW. 1970. Accumulation of organic
matter in a chronosequence of soils developed on wind-blown
sand in New Zealand. J Soil Sci 21:146–53.
Tackett NW, Craft CB. 2010. Ecosystem development on a
coastal barrier island dune chronosequence. J Coast Res
26:736–42.
Tapsell P, Newsome D, Bastian L. 2003. Origin of yellow sand
from Tamala Limestone on the Swan Coastal Plain, Western
Australia. Aust J Earth Sci 50:331–42.
Tejan-Kella MS, Chittleborough DJ, Fitzpatrick RW, Thompson
CH, Prescott JR, Hutton JT. 1990. Thermoluminescence dating
of coastal sand dunes at Cooloola and North Stradbroke Is-
land, Australia. Aust J Soil Res 28:465–81.
Thompson CH. 1981. Podzol chronosequence on coastal dunes
of eastern Australia. Nature 291:59–61.
Thompson CH. 1992. Genesis of podzols on coastal dunes in
Southern Queensland: I. Field relationships and profile mor-
phology. Aust J Soil Res 30:593–613.
Turner BL, Condron LM. 2013. Pedogenesis, nutrient dynamics,
and ecosystem development: the legacy of T.W. Walker and
J.K. Syers. Plant Soil 367:1–10.
Turner BL, Romero TE. 2009. Short-term changes in extractable
inorganic nutrients during storage of tropical rain forest soils.
Soil Sci Soc Am J 73:1972–9.
Turner BL, Condron LM, Richardson SJ, Peltzer DA, Allison VJ.
2007. Soil organic phosphorus transformations during pedo-
genesis. Ecosystems 10:1166–81.
Turner BL, Condron LM, Wells A, Andersen KM. 2012. Soil
nutrient dynamics during podzol development under lowland
temperate rain forest. Catena 97:50–62.
Vitousek PM. 2004. Nutrient cycling and limitation. Princeton:
Princeton University Press.
Vitousek PM, Farrington H. 1997. Nutrient limitation and soil
development: experimental test of a biogeochemical theory.
Biogeochemistry 37:63–75.
Walker TW, Syers JK. 1976. The fate of phosphorus during
pedogenesis. Geoderma 15:1–19.
Walker J, Thompson CH, Fergus IF, Tunstall BR. 1981. Plant
succession and soil development in coastal sand dunes of
subtropical eastern Australia. In: West DC, Shugart HH, Bot-
kin DB, Eds. Forest succession: concepts and application. New
York: Springer. p 107–31.
Walker LR, Wardle DA, Bardgett RD, Clarkson BD. 2010. The
use of chronosequences in studies of ecological succession and
soil development. J Ecol 98:725–36.
Wardle DA, Walker LR, Bardgett RD. 2004. Ecosystem properties
and forest decline in contrasting long-term chronosequences.
Science 305:509–13.
Wardle DA, Bardgett RD, Walker LR, Peltzer DA, Lagerstro¨m A.
2008. The response of plant diversity to ecosystem retrogres-
sion: evidence from contrasting long-term chronosequences.
Oikos 117:93–103.
Wells A, Goff J. 2007. Coastal dunes in Westland, New Zealand,
provide a record of paleoseismic activity on the Alpine fault.
Geology 35:731–4.
White PJ. 2012. Long-distance transport in the xylem and
phloem. In: Marschner P, Ed. Marschner’s mineral nutrition
of higher plants. London: Academic Press. p 49–70.
Williamson WM, Wardle DA, Yeates GW. 2005. Changes in soil
microbial and nematode communities during ecosystem de-
cline across a long-term chronosequence. Soil Biol Biochem
37:1289–301.
Woods PJ, Searle DJ. 1983. Radiocarbon dating and Holocene
history of the Becher/Rockingham Beach ridge Plain, West
Coast, Western Australia. Search 14:44–6.
Wyrwoll KH, King PD. 1984. A criticism of the proposed regional
extent of Late Cenozoic arid zone advances into south-wes-
tern Australia. Catena 11:273–88.
Wyrwoll K-H, Turner BL, Findlater P. 2014. On the origins,
geomorphology and soils on the sandplains of south-western
Australia. In: Lambers H, Ed. Plant life on the sandplains in
Southwest Australia, a global biodiversity hotspot. Crawley:
University of Western Australia Publishing. p 3–23.
Soil Development and Nutrient Availability
Author's personal copy