Professional Documents
Culture Documents
Eduardo Garzanti
PII: S0012-8252(18)30606-8
DOI: https://doi.org/10.1016/j.earscirev.2018.12.014
Reference: EARTH 2754
To appear in: Earth-Science Reviews
Received date: 18 October 2018
Revised date: 16 December 2018
Accepted date: 18 December 2018
Please cite this article as: Eduardo Garzanti , Petrographic classification of sand and
sandstone. Earth (2018), https://doi.org/10.1016/j.earscirev.2018.12.014
This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.
ACCEPTED MANUSCRIPT
1
PT
Milano-Bicocca, Milano (Italy)
Tel: +39 02 64482088, Email:
RI
SC
NU
ABSTRACT
Descriptive petrographic classifications of sand and sandstone proposed more than half a century
MA
ago are still in use, although they were formulated at a time when depositional and post-depositional
sedimentary processes were poorly understood, and before the relationships between tectonics and
ED
scientific articles and technical reports are still encumbered with obsolete concepts, graphical tools,
T
and ambiguous terminology that make sediment descriptions awkward and misleading. A
EP
The descriptive petrographic classification of sand and sandstone proposed in this paper is based on
AC
the quasi-universally used Gazzi-Dickinson point-counting method, and simply translates into
words ternary compositions of quartz, feldspar, and lithic fragments without introducing any new
names. The classic QFL plot is subdivided into 15 fields - labelled by adjectives introduced long
ago by K.A.W. Crook and endorsed by W.R. Dickinson and more recently by G.J. Weltje - which
reflect relative abundances of the three main framework components (provided they exceed
10%QFL). According to standard use, the less abundant component goes first, the more abundant
last (e.g., litho-feldspatho-quartzose composition translates into Q > F > L >10%QFL). For lithic-
rich sand and sandstone, information on the prevailing rock fragment type can be added by an
ACCEPTED MANUSCRIPT
2
additional free adjective (e.g., metamorphiclastic, carbonaticlastic), as proposed long ago by R.V.
Ingersoll. For lithic-poor feldspatho-quartzose and quartzose sand and sandstone, further formal
subdivisions are proposed, thus reaching a total of 18 compositional fields overall. Modern sand
known to be derived from different source rocks and found in major world’s rivers, deserts, and
deep-sea fans fits in the pigeonholes defined by the relative abundance of quartz, feldspar, and lithic
fragments.
PT
The aim of this classification is to restore directness in sandstone petrology, and to avoid
ambiguities generated in the past by making reference to badly defined archetypes, such as
RI
greywacke or arkose, thus confusing petrographic composition with subjective considerations about
SC
plate-tectonic setting, texture, hydraulic behaviour, mechanical durability, or chemical durability in
the illusion that a classification could be genetic at the same time as descriptive.
NU
“You ask what is the use of classification, arrangement, systematization?
I answer you: order and simplification are the first steps toward the
mastery of a subject — the actual enemy is the unknown.”
MA
Keywords
Sandstone petrology; Gazzi-Dickinson method; QFL diagram; Sediment provenance; Plate-tectonic
T
setting; Texture and composition; Greywacke and arkose; Rock fragments; Chert, carbonate, and
EP
1. INTRODUCTION
C
AC
The petrographic study of sedimentary archives is one of the many keys to deciphering geological
history. The road is, however, long and winding, and a firm methodological approach is essential to
staying on track. This article reviews operational and conceptual flaws of traditional classification
criteria and suggests how to describe and classify in a clear and exhaustive way the composition of
sand or sandstone samples in a scientific article or technical report. The straightforward scheme
proposed here which focuses strictly on petrographic composition, expressly neglecting texture,
ACCEPTED MANUSCRIPT
3
started to flourish in the first half of the twentieth century, and culminated with sandstone
classifications proposed in numbers between the late 1940s and the early 1960s (as summarized in
Klein, 1960; Okada, 1971; Scholle, 1979). These classification schemes (Fig. 1; Fig. 2) inevitably
PT
suffered from the level of information available at the time about sedimentary and geodynamic
RI
processes. Investigations on diagenetic transformations were limited, and the nature of interstitial
SC
phyllosilicates in sandstone consequently poorly understood. Before advent of the plate-tectonic
theory, relationships between tectonics and sedimentation were still immersed in a panorama full of
NU
mythological entities, including the large geosyncline clan (Krynine, 1948; Kay, 1951; Folk, 1968).
Irruption of the new geodynamic paradigm in the 1960s and 1970s revolutionized the petrological
MA
study of terrigenous rocks as well, leading to novel ideas embodied in the genetic sandstone-
classification scheme of Dickinson (1985; Dickinson and Suczek, 1979) that has dominated the
ED
scene since then. The enthusiasm generated by such a major conceptual step forward, however,
created the illusion that sediments generated in different plate-tectonic settings should inevitably
T
EP
plot in separate fields within a QFL diagram. The uncritical use of such a simple graphic tool as a
years (e.g., Mack, 1984; Molinaroli et al., 1991; Weltje, 2006), ending up as a ready-made approach
AC
This conservative attitude has persisted. Even in recent scientific articles, it is common to find
such as arkose or greywacke, the use of which has been contested since their early introduction two
centuries ago. In too many cases, genetic interpretations are still based on the belief that numerical
parameters readily obtained by petrographic analysis may open an easy way to the reconstruction of
past (Pettijohn, 1948 p.113; Rodgers, 1950 p.299) is that a classification could, and even should be
at the same time both descriptive and genetic (i.e., objective and subjective). Sediment mineralogy,
however, reflects the multiple superposed effects of numerous controlling factors, including source-
rock lithology, climate, and tectonic activity, together with diverse physical and chemical processes
affecting detritus through one or more sedimentary cycles. Complex equations with many
PT
2. WHY A CLASSIFICATION SHOULD NOT BE GENETIC
RI
SC
Whenever we find ourselves baffled by the variety and complexity of nature’s products and
phenomena, we may start to make comparisons, look for similarities and differences, and finally
NU
combat chaos by dividing objects into categories and giving them a name. A classification is a
simple artificial means to impose order upon the real world. This process leads to the formulation of
MA
a language that, when widely agreed upon, allows us to exchange information. Words, however, do
not only technically describe objects, but also create suggestions. Not differently from myth and
ED
religion, scientific theories grow with new words, the spell of which may help to conceal the
T
unbridgeable distance separating our models from truth. In the geosciences, evocative words have
EP
always been used to depict processes and scenarios which, once meeting wide consensus, proved
easy to believe and difficult to depart from (Dickinson, 2003). In order to make diligent use of
C
AC
reason, digest observations, and progress on the infinite stairway to knowledge, however, we need
to focus on naked phenomena, stripped of rhetoric, superposed scent of incense, and special effects
of any kind. Sharp reasoning needs clear words and concepts devoid of ambiguities and implicit
assumptions.
A classification, therefore, should be no more than a basic tool, an attempt to communicate reality
in a direct way. An efficient descriptive language cannot substitute for understanding, but may
represent the first step on the path toward understanding. Hoping that the goal can be reached
without making the journey is laziness. Believing that enlightenment may come in a moment is
ACCEPTED MANUSCRIPT
5
illusory. The commandment “genesis must and does permeate our classification” (Pettijohn, 1948
p.113) or the promise that plotting a point into a QFL diagram is sufficient to reveal a geodynamic
context (Dickinson and Suczek, 1979) are traps into which we should not fall.
PT
Plate tectonics is the best paradigm at our disposal to describe geological processes at planetary
scale. This does not mean that full understanding has been reached, and several key questions
RI
remain unanswered about the processes that shape the face of our planet. We are still debating when
SC
and how plate tectonics began on Earth, why plates are moving, what drives oceanic and continental
subduction, whether plates breakoff at depth, what forces cause subsidence in orogen-related basins,
NU
or what triggers gigantic outbursts of magmas (e.g., Hamilton, 2011; Korenaga, 2013; Doglioni and
MA
Panza, 2015; Garzanti et al., 2017a). Tectonic processes are reflected in sediment composition, and
sedimentary petrology thus represents one fruitful way to extrapolate knowledge acquired in
ED
modern settings to reconstruct tectonic evolution throughout the geological past. Using sediment
composition as a key to paleogeodynamics is the fundamental idea underlying the work of W.R.
T
Dickinson and co-workers (e.g., Dickinson and Suczek, 1979; Ingersoll and Suczek, 1979), an idea
EP
that stands as valid as ever. Significant parts of the operative procedure, however, need to be
C
revised.
AC
As shown by Molinaroli et al. (1991) and Weltje (2006), the uncritical use of Dickinson’s plots is
bound to meet with limited success, and not only because relevant factors such as grain-size control,
sampling scale, and environmental or diagenetic bias are neglected (Ingersoll, 1990; Johnsson,
1993; Weltje and von Eynatten, 2004; Basu, 2017) but also and more fundamentally because some
implicit assumptions about sediment generation embodied in this model are invalid. These include
the tenet that vast areas covered by continental flood basalts cannot supply large amounts of detritus
transcontinental river systems such as the Nile, the Orange, or to a lesser extent the Yang Tze
(Garzanti et al., 2006a, 2012; Vezzoli et al., 2016). Flawed is also the hope that petrographic
analysis alone could discriminate between detritus shed from neometamorphic domains forming the
axial core of young mountain belts versus paleometamorphic detritus derived from old orogenic
roots exposed in cratonic shields or uplifted on the shoulders of continental rifts. In the lack of
information from detrital geochronology and geochemistry, orogenic sandstone shed from thrust
PT
belts and magmatic arcs cannot be safely distinguished from anorogenic sandstone derived from
RI
continental blocks and flood basalts (e,g., Garzanti et al., 2015b). Framework petrography cannot
SC
tell us whether source rocks are allochthonous or autochthonous. Therefore, orogenic and
2016).
MA
Texture and composition are independent variables. Texture of clastic sediments is primarily
T
detail, i.e., average (mean, median, mode), uniformity (sorting), asymmetry (skewness), and
C
peakedness (kurtosis; Folk, 1966), whereas composition is defined by the relative percentages of
AC
framework components, i.e., single minerals and rock fragments. The repeated attempts to combine
in a single classification scheme compositional, textural, and even hydraulic properties at the same
time (provenance, maturity, and fluidity factors of Pettijohn, 1954) has created much confusion
since the dawn of sandstone petrography (Table 1; Klein, 1963 p.569-570) and still reverberates in
The emblematic example is the archaic term greywacke, used originally by miners to describe hand
specimens and successively adopted in the geological literature to indicate a generic lithic-rich
ACCEPTED MANUSCRIPT
7
composition as well as colour (grey), texture (poorly sorted, supposedly matrix-rich), and
depositional environment (turbidite). The crux of ambiguity stems from putting composition and
texture in the same basket (Fig. 2), often implying that “immature” composition (i.e., low quartz)
should necessarily combine with “immature” texture (i.e., poor sorting). For instance, in Pettijohn
achievement of textural maturity”. Many sandstone petrographers thus shared G.H. Packham’s view
PT
(1954) that a “greywacke suite” deposited by turbidity currents could be distinguished from an
RI
“arkose-quartzose sandstone suite” deposited by traction currents (McBride 1963; Crook 1974), a
dichotomy that has no basis (Okada, 1966; Garzanti, 2017). The tangle tightened since “clay
SC
matrix” was chosen as an end-member of sandstone composition in most formal classification
NU
schemes proposed in the 1950s (e.g., Pettijohn 1949; Dapples et al. 1953; Gilbert 1954; Packham
1954; Bokman 1955; Crook 1960). Later on, several authors emphasized the rarity of modern sand
MA
characterized by very poor sorting (Cummins, 1962; Whetten, 1969), and concluded that the
supposedly peculiar aspect of greywackes was not original, but produced during diagenesis by a
ED
variety of processes including plastic deformation of altered volcanic rock fragments or other soft
clasts of either extrabasinal or intrabasinal origin to create what has been termed pseudomatrix
T
EP
(Dickinson, 1970; Whetten and Hawkins, 1970; Galloway, 1974). The “greywacke problem” was
solved.
C
AC
The term greywacke (grauwacke in German, graywacke in American English) was formally
introduced in 1785 by mine director F.W.H. von Trebra, adopted by A.G. Werner (1787 p.18) and
defined as “quartz breccias with flakes of micas and fragments of chert or sandstones in a cement of
clay” (Lasius, 1789). The type greywackes are upper Paleozoic deep-water sandstones of the Hartz
mountains that contain quartz as well as abundant feldspar and diverse plutonic, volcanic,
metamorphic, and sedimentary rock fragments (Helmbold and van Houten, 1958; Huckenholz,
ACCEPTED MANUSCRIPT
8
1963). The term has generated confusion and has been harshly criticized since its early introduction
(Mawe, 1818 p.92; Sedgwick and Murchison, 1839 p.260; Murchison, 1854 p.359; Krynine, 1956;
Boswell, 1960; Okada, 1971), until Folk (1968 p.128) conclusively pointed out that greywacke is
nothing else than “a very hard, ugly, dirty, dark rock that you can't tell much about in the field”.
PT
The term arkose was introduced by Brongniart (1826) and originally defined as sandstone
RI
containing either more quartz than feldspar (arkose commune) or more feldspar than quartz (arkose
SC
granitoide). The type arkose derived from the Massif Central in France, however, may contain a
variety of igneous (granite), sedimentary (sandstone), and metamorphic (quartzite) rock fragments,
NU
largely overlapping the type greywacke in both texture and composition (Huckenholz, 1963; Dott,
1964 p.626; Dickinson, 1970 p.697). As for greywacke, sub-greywacke, high-rank greywacke and
MA
low-rank greywacke, the compositions of arkose (or “arkosite”), impure arkose, arkosic arenite,
arkosic wacke, or subarkose have been defined through time with inconsistent boundaries (Fig. 1;
ED
Fig. 2; Oriel, 1949; Scholle, 1979). Consequently, arkose and subarkose are at best imprecise
T
sandstones. As R.H. Dott (1964 p.625) put it: “the name arkose itself has little descriptive merit;
derived from any type of source rock, garbage cans that potentially host any kind of extrabasinal
and intrabasinal, larger and smaller, platy and equant, low-density and high-density, carbonate and
By the simple procedure presented here, a composite descriptive label is attached to sand or
sandstone samples based on their petrographic composition (Fig. 3). Because communicative
efficiency makes a scientific article or technical report shorter and clearer, the aim is to transmit the
richest information with the fewest words. If not widely agreed upon and adopted, however, a new
terminology only introduces unwanted noise. Taking advantage of already existing methods and
familiar terms, provided they are appropriate, is sensible, economic, and may facilitate acceptance
PT
by the community. For this reason, the proposed classification: a) is based on a consolidated
RI
analytical technique (i.e., Gazzi-Dickinson method; Ingersoll et al., 1984; Zuffa, 1985); b) makes
SC
use of fundamental parameters and of a universally recognized graphical display (i.e., QFL plot); c)
Gazzi (1966 p.73-74) started from the obvious consideration that coarse-grained rock fragments
ED
comparable quantitative data from samples of different grain size, he proposed that minerals
EP
occurring within rock fragments and exceeding 30 m in size – the conventional boundary between
frictional grains and cohesive matrix as well as the thickness of a standard thin section (Spencer,
C
1963; Dott, 1964 p.630-631) – should be reunited in the dataset with single detrital minerals of the
AC
same type.
A similar, albeit simplified version of this operational procedure was proposed independently a few
years later by Dickinson (1970), who distinguished formally between aphanite lithic fragments (L)
to be counted as such, and microphanerite rock fragments (R) to be assigned according to the
mineral beneath the cross-hair. The boundary between fine-grained aphanites and coarse-grained
phanerites was chosen instead as 62.5 m (i.e., the lower limit of the sand range), as agreed by P.
Gazzi’s former student G.G. Zuffa (1980 p.27, 1985). As a partial solution to this problem in case
ACCEPTED MANUSCRIPT
10
of data collection with the traditional QFR method, in which rock fragments are counted as such,
(Suttner et al., 1981), Folk (1980 p.127) proposed to group coarse-grained granitoid and granitoid-
gneiss grains with feldspar in pole F, and to assign only fine-grained volcanic, metamorphic, and
The Gazzi-Dickinson method was eventually formalized by Ingersoll et al. (1984) and successively
widely accepted by most sandstone petrographers under the belief that “use of the method minimizes
PT
variation of composition with grain size, thus eliminating the need for sieving and multiple counts
RI
of different size fractions”, which is overly optimistic (Garzanti et al., 2006a p.331). The rectified
statement found a dozen lines below in Ingersoll et al. (1984 p.103) reads: “there are two reasons
SC
for variation of modal composition with grain size: 1) the breakage of fragments into constituent
NU
grains, and 2) actual mineralogic variation with grain size. The Gazzi-Dickinson method
A consensus has long been reached among sedimentary petrographers that, as a first approximation,
EP
sand and sandstone can be considered as ternary mixtures of three principal components: quartz (Q;
most common mineral species in the Earth’s crust), feldspars (F; most common mineral group in the
C
Other components are generally less abundant, considered as accessories, and neglected for
classification purposes not only to reduce complexity but also to come closer to a transport-
invariant measure of sediment composition (Weltje, 2004). Most accessories have peculiar
hydraulic behaviour because of their distinct shapes (e.g., slow-settling platy phyllosilicates) or
densities (e.g., fast-settling dense minerals), and can be strongly concentrated or depleted locally in
(Garzanti et al., 2008, 2009). Intrabasinal grains, which in sediment samples may be mixed in any
ACCEPTED MANUSCRIPT
11
proportion with terrigenous extrabasinal detritus derived from erosion of source rocks, must be
considered independently for classification purposes (fig. 3 in Zuffa, 1980; fig. 1 in Garzanti, 1991).
Once major framework components are reduced to three, sand and sandstone can be readily
classified making use of the classical QFL triangular diagram, in which each point represents a
ternary composition obtained by the Gazzi-Dickinson point-counting method. A most precise way
PT
among the three major components (e.g., with a format like Q 45 F18 L37 ; Dickinson, 1970). Although
RI
appropriate for tabulated data, however, an analytical script such as this singles out each object of
SC
the suite, and a classification of singles is no classification at all.
NU
5.3. Nomenclature
Most traditional classifications of sand and sandstone are based on the quartz-feldspar-lithics (QFL)
MA
or quartz-feldspar-rock fragments (QFR) triangle, subdivided into several fields (generally 5 to 10),
each labelled differently and delimited by necessarily arbitrary conventional boundaries (Fig.1; Fig.
ED
2). The first step of the procedure proposed here is the same as in Weltje (2006 p.82): by tracing
“three lines from each of the vertices towards the middle of the opposite sides, i.e. lines along which
T
EP
the abundance of one component equals that of another” the QFL space is subdivided in 6 equal
feldspatho-quartzose, and quartzo-feldspathic (Fig. 3B). The adjectives feldspathic (F/L > 3), litho-
AC
feldspathic (3 > F/L > 1), feldspatho-lithic (3 > L/F > 1) and lithic (L/F > 3) were originally
proposed by Crook (1960 p.425), and considered by Dickinson (1970 p.697) to “permit adequate
G.J. Weltje’s straightforward classification, however, considers only the ratio between two of the
three major components, and conveys no information on the abundance of the third. Moreover, very
different compositions (e.g., Q 99 F1 L0 and Q 35 F33 L32 ) would plot in the same field. Additional fields
are needed to increase classification efficiency and discrimination power. A simple effective way is
ACCEPTED MANUSCRIPT
12
to trace another three lines corresponding to a 10% relative content of each major component, thus
obtaining another 9 fields (15 overall; Fig. 3C). The six central fields, where all three major
fields along the three legs of the QFL triangle, where one major component does not exceed 10%,
PT
are labelled as litho-feldspathic (LF), feldspatho-lithic (FL), quartzo-lithic (QL), litho-quartzose
RI
(LQ), feldspatho-quartzose (FQ), and quartzo-feldspathic (QF). The three rhomboidal fields at the
SC
apices, where two out of the three major components do not exceed 10%, are thus simply labelled
lithic (L), feldspathic (F), and quartzose (Q). The rationale here is that a component below 10% can
NU
be ignored for nomenclatural purposes.
These 15 fields are sufficiently narrow to be meaningful and at the same time sufficiently wide to
MA
allow samples to be classified by a careful inspection under the microscope, even without the need
of full quantitative analysis. Difficulties may arise for samples in which quartz, feldspar, and lithic
ED
fragments are all present in subequal proportions (i.e., ≥ 30%QFL), in which case a QFL field
may be informally considered. Observations of sand and sandstone specimens with a hand lens can
T
EP
hardly be accurate enough to apply this classification in the field, where more generic terms such as
In conformity with the original use by Crook (1960) and Dickinson (1970, 1985), and differently
AC
from Weltje, 2006), the prevalent component goes last, so that a litho-feldspatho-quartzose sand has
more quartz than feldspar, and more feldspar than lithics. The main drawback with this
the other hand, these composite names add to the precision of petrographic descriptions and are
immediately intelligible in terms of relative composition. This economic scheme does not require
the introduction of new fancy names and gets rid of the use of reference standards to define the
composition of rocks (e.g., type greywacke or type arkose; Huckenholz, 1963), an “archetype”
ACCEPTED MANUSCRIPT
13
doctrine already superseded in many fields of the geosciences, from igneous petrology to
When detrital components in sand or sandstone samples are highly diverse and potentially so
PT
numerous, can a classification based on three main framework components only be satisfactorily
complete? The answer to this rhetorical question is that two further specifications are needed to
RI
make the classification system sufficiently informative. For lithic-rich sand and sandstone we may
SC
designate the dominant group (e.g., volcanic, plutonic, metamorphic, or sedimentary) and possibly
specific type (e.g., felsic or mafic, low-grade or high-grade, gneiss or sepentineschist, limestone or
NU
chert) of rock fragments, which are the carriers of most robust provenance information. For lithic-
MA
poor sand and sandstone, instead, we may specify the relative abundance of quartz and feldspar, and
possibly the dominant feldspar type as well (i.e., K or Fk = K-feldspar vs. P or Fp = plagioclase).
ED
Adding independent textural information is also useful, because the relative abundance of quartz,
feldspar, and lithic fragments may be markedly influenced by sample grain size.
T
EP
If lithic fragments are few, then it is essential to retrieve and incorporate in the classificatory label
AC
what quartz and feldspars may reveal. Anorogenic sediments deposited along passive margins and
fed by rivers draining continental interiors typically consist of quartz and feldspar mostly
(continental block provenance of Dickinson, 1985; Sciunnach and Garzanti, 2012). In such cases,
key information is provided by the Q/F ratio, traditionally considered as controlled by the
competition between chemical weathering and unroofing rate in source areas. Feldspar prevails
where granitoid basement is unroofed either very rapidly (tectonic arkose of Folk, 1980; basement
uplift sub-provenance of Dickinson, 1985) or in arid climate (climatic arkose of Folk, 1980),
ACCEPTED MANUSCRIPT
14
whereas quartz dominates in case of intense weathering or extensive recycling of older quartz-rich
sandstone in low-relief shield areas during long periods of tectonic quiescence (craton interior sub-
The necessity to formally distinguish between sand and sandstone with Q/F < 1 and typically P > K
(ideal arkose of Dickinson, 1985) from those with Q/F > 1 and commonly P < K (plagioclase being
widely considered more weatherable than alkali feldspar; Goldich, 1938; Velbel, 1993) has been
PT
felt since Brongniart (1826) formally distinguished between arkose granitoide (Q/F < 1) and arkose
RI
commune (Q/F > 1). Different definitions contemplating diverse conventional boundaries have been
SC
proposed since then. In the classification of Folk (1980 p.127), an arkose contains less than 75%
quartz and more than 18.75% feldspar (Q/F < 4), a subarkose from 75% to 95% quartz and from
NU
2.5% to 25% feldspar (3 < Q/F < 38), and a quartzarenite more than 95% quartz (Q/F > 19).
The scheme proposed here to classify lithic-poor sand and sandstone, particularly fruitful in the
MA
study of modern passive-margin sand generated in the subequatorial climatic belt (Garzanti et al.,
2018a), is based on the Q/F ratio and allows identification of diverse categories required to convey
ED
feldspar-rich subclass (fFQ; 1 < Q/F < 2), which may be differentiated further into plagioclase-rich
T
EP
if P/K > 2 (pFQ) or K-feldspar-rich if K/P > 2 (kFQ), and a quartz-rich subclass (qFQ; 4 < Q/F <
9). Distinguished within the quartzose class is a pure quartzose subclass (qQ; Q%QFL > 95),
C
corresponding to the quartzarenite of Folk (1980). The complete classification scheme including
AC
these refined subdivisions consists of 18 compositional fields overall (Fig. 3D), allowing us to
capture in one composite name the essential petrographic features of sand and sandstone.
The description of lithic-rich samples poses the opposite problem: an excess of information that can
be hardly compressed in the classificatory label. The challenge was tackled by Folk (1968), who
increased the informative value of his classification by adding auxiliary triangles to the main QFR
ACCEPTED MANUSCRIPT
15
diagram (Fig. 1). His MRF-VRF-SRF ternary plot formally distinguishes among lithic-rich
named phyllarenite, volcanic arenite, and sedarenite – and next among sedarenites containing
mainly carbonate, chert, or terrigenous rock fragments – named calclithite, chert-arenite, and shale-
arenite or sandstone-arenite –. Ingersoll and Suczek (1979) introduced the LmLvLs ternary plot, the
PT
badly wherever detrital components cannot be reduced to three groups sharply, and MRF-VRF-SRF
RI
or LmLvLs plots suffer from at least two major problems. First, the boundary between sedimentary
SC
and metasedimentary grains, as well as between volcanic and metavolcanic grains, is difficult to
define in order to obtain consistent results by different operators (Wolf, 1971; Dickinson, 1985
NU
p.338). Progressively more detailed operational solutions have been suggested in subsequent years
(fig. 4 in Dorsey, 1988; fig. 7 in White et al., 2002; figs. 1 to 4 in Garzanti and Vezzoli, 2003), but
MA
the problem remains, for instance, in the distinction between sparitic limestone or dolostone versus
marble grains. Secondly, grains that carry major provenance information such as ultramafic or fine-
ED
grained plutonic clasts (e.g., granophyre) are not unequivocally assigned. All igneous grains may be
grouped within pole Lv, perhaps including cellular serpentinite (Fig. 4O), whereas foliated
T
EP
serpentineschist (Fig. 4P) may join pole Lm with undetermined ultramafic grains split 50-50%.
Other operational choices, including how to consider chert and carbonate grains, are tackled in
C
section 7, but the main point here is that the spectrum of rock fragments is so wide that any attempt
AC
to handle them with a rigid procedure is doomed to failure in complex natural cases. A flexible
As a simple way to indicate prevalent types of rock fragments (reference to rock fragments rather
than to lithics is preferred here to maximize provenance information), the adjectives plutoniclastic,
widely used volcaniclastic. These terms, introduced by Ingersoll (1983 p.1137), were originally
meant to be “genetic and interpretive, as opposed to descriptive”. However, in conformity with the
ACCEPTED MANUSCRIPT
16
leitmotif of this article expressed in section 2, these adjectives may be better used to inform
objectively about the dominant rock-fragment type in the sample without any subjective genetic
For purely volcaniclastic sediments or sedimentary rocks, names such as rhyodacitic and andesitic
sand (Dickinson, 1985) or rhyolite-arenite and andesite-arenite (Folk, 1980 p.128) have been
PT
proposed. In a similar way, the diverse end-member compositions displayed by modern sand
RI
derived from monolithologic sources within distinct tectono-stratigraphic levels of a lithospheric
SC
section (first-order sampling scale of Ingersoll et al, 1993) were designated as metarhyodacite sand,
slate sand, phyllite sand, schist sand, gneiss sand, kinzigite sand, stronalite sand, granite sand,
NU
gabbro sand, and peridotite sand (Garzanti et al., 2006b). In the general situation of polylithologic
sources and mixed rock-fragment types, none of which dominant, terms such as chert-bearing or
MA
component. The guideline of this articulated free scheme is to borrow previously proposed criteria
ED
and terms to devise a coherent procedure by which every sample of sand or sandstone can be given
Whereas single minerals may derive from diverse rock types, the texture and mineralogy of rock
fragments generally point uniquely to a specific lithology (Fig. 4). Such information is invaluable in
provenance analysis and must not be lost during data collection when using the Gazzi-Dickinson
point-counting method (Gazzi and Zuffa, 1970; Suttner and Basu 1985). As recommended by Gazzi
(1966), this can be done either by using a detailed point-counting sheet that allows the simultaneous
registration of the mineral beneath the cross-hair and of the rock fragment in which the mineral is
located (Fig. 5; table 1 in Zuffa, 1980; table 3 in Zuffa, 1985; table 2 in Fontana et al., 1985), or by
ACCEPTED MANUSCRIPT
17
dedicating a second count specifically to rock-fragment types (Ingersoll and Suczek, 1979 p.1220;
The effort to discriminate in great detail among the vast spectrum of rock fragments potentially
encountered in sand and sandstone (Fig. 4) in order to retain the maximum possible level of
PT
information may lead to an unacceptable degree of inhomogeneity in datasets. A robust
RI
classification of rock fragments is thus necessary to ensure comparability among results obtained by
SC
different operators.
hornfelsic). Building on Graham et al. (1976), Ingersoll and Suczek (1979 p.1220-1221) introduced
ED
additional categories for metamorphic lithics based on their mineralogy and texture (i.e.,
T
metamorphic rank into Lm1 (slate, quartzite, slatey siltstone) and Lm2 categories (phyllite-schist,
C
a further refinement, White et al. (2002) classified metasedimentary lithics according to both
protolith (pelitic vs. felsitic) and rank (Lm1 = slate, metasandstone; Lm2 = phyllite; Lm3 =
micaschist, gneiss). Garzanti et al. (2002a) emphasized the importance of ultramafic detritus,
serpentineschist grains (Fig. 4 O,P). Garzanti and Vezzoli (2003) proposed a comprehensive
operational classification scheme, illustrated by 24 prototypical grains, that considers four protolith
compositions (pelitic, psammitic or felsic volcanic, carbonatic, mafic volcanic) and six successive
ACCEPTED MANUSCRIPT
18
suite of a hundred rock-fragment types and subtypes including sophisticated categories such as
“amphibole-bearing quartz-feldspar gneiss of rank 4” can be used consistently during routine point-
This does not mean that identification of rock fragments under the microscope is easy, especially in
strongly diagenetically-altered sandstone samples, where even grain boundaries are hard to
PT
establish. Long-standing problems include difficulties in distinguishing felsitic volcanic rock
RI
fragments from impure unfossiliferous chert (Wolf, 1971), correct identification of pseudomatrix as
SC
groundmass generated by the alteration and deformation of lithic grains (Dickinson, 1970), and the
troublesome distinction between terrigenous extrabasinal versus intrabasinal carbonate and non-
NU
carbonate grains (Zuffa, 1980, 1985; Garzanti, 1991). An analogous thorny issue, represented by the
differentiation of neovolcanic versus paleovolcanic detritus (i.e., grains originated from erosion of
MA
penecontemporaneous vs. notably older volcanic rocks; Zuffa, 1980; Critelli and Ingersoll, 1995),
does not generate problems in the classification of sand and sandstone because coeval and non-
ED
coeval volcanic lithics are all assigned to pole L. This distinction, based on criteria described in
There is no general consensus on how these problems should be treated operationally, whether we
should surrender to challenging grain identifications or not, and whether lithic grains should be
C
grouped by their origin or by their presumed stability. Idealized concepts such as “stability” or
AC
“maturity” are highly questionable (Garzanti, 2017), because grains may be durable in certain
environments but labile in others. Most detrital minerals including olivine and pyroxene can resist
sediment transport over thousands of kilometers in high-energy fluvial, eolian, and marine
environments (Garzanti et al., 2015a, 2015b), whereas none, including quartz and zircon, can be
diagenesis (Crook, 1968; Cleary and Conolly, 1970; Garzanti et al., 2013a, 2018b). Clarity and
objectiveness, which are the essential requirements of a classification, are lost when subjective
ACCEPTED MANUSCRIPT
19
assumptions are made concerning the relative durability of grain types independently of specific
processes and physico-chemical conditions encountered during erosion, transport, deposition, and
burial.
By giving privilege to stability, Gazzi (1966 p.74) and next Dickinson (1970 p.696) chose to group
PT
in pole Q “the sum of quartz and chalcedony grains of all kinds“. Following Krynine (1948),
initially also Folk (1954) decided to assign chert to pole Q, but next changed his mind in favour of
RI
pole R (Folk, 1980 p.127), as in van Andel (1958), Füchtbauer (1959), Chen (1968), and Okada
SC
(1971) (Table 1). As a conciliatory solution, Dickinson and Suczek (1979) included both
polycrystalline quartz and chert in sub-pole Qp, grouped with monocrystalline quartz in the QtFL
NU
plot, and with lithics in the QmFLt plot (Fig. 1). Recurrence to these two different versions of the
MA
QFL diagram is not straightforward and, because of different field boundaries, may make a
dependent on grain size and/or on the convention adopted during point-counting, that chert and
felsic volcanic rock fragments are not invariably distinguishable, and that gradation from pure chert
T
to argillite is also common (Folk, 1980 p.126; Dickinson, 1985 p.336-337), the easiest and most
EP
reasonable escape from the tangle is to group all chert grains to pole L (Zuffa, 1980; Dorsey, 1988)
C
and all quartz grains to pole Q (Fontana et al., 1985). This simple procedure allows using a single
AC
QFL diagram for classification with great gain in simplicity and clarity.
Difficulties in identifying the intrabasinal versus extrabasinal origin of carbonate grains, and of
sedimentary grains in general, has long been acknowledged (e.g., Gazzi, 1966 p.76). Zuffa (1980,
1985) provided detailed operational criteria to discriminate intraclasts, bioclasts, ooids, or peloids
generated within the sedimentary basin and coeval with deposition versus limestone or dolostone
ACCEPTED MANUSCRIPT
20
rock fragments derived from erosion of older carbonate rocks outside of the sedimentary basin (Fig.
4A,B). Because such distinction is seldom straightforward, grains that cannot be recognized with
category (Wolf, 1965; Blatt et al., 1972 p.460; Zuffa, 1980 p.26). Limeclasts may be next
tentatively reapportioned to either the intrabasinal or extrabasinal group, in the worst case by a 50-
PT
The problem is even more complicated because non-coeval intrabasinal grains exist as well, such as
RI
those reworked locally from previously deposited layers affected by early cementation or
SC
pedogenesis (e.g., beach rock, eolianites, caliche), or eroded from underlying depositional
sequences during episodes of tectonic uplift or eustatic lowstand (e.g., Garzanti et al., 2003; 2017b).
NU
Such complexities may regard a wide variety of grains, including clay chips, gypsum, glaucony,
chert, or phosphate clasts reworked from mudcracks, soil profiles, sabkhas, or marine hardgrounds
MA
Problems can be tackled, running the risk of making errors, or skipped. On the one side, Mack
T
EP
(1984 p.218) concluded that “carbonate rock fragments provide important information about
source rock and probably should be included in detrital modes”. On the other side, Dickinson (1985
C
p.336) chose not to recalculate extrabasinal carbonate grains with other lithic fragments “because of
AC
their vastly different geochemical response during weathering and diagenesis, as well as the ease of
confusion with intrabasinal carbonate grains”. Using durability as a criterion for discriminating
certain grain types, however, spoils any attempt to design a descriptive classification by introducing
climatic conditions (Gazzi et al., 1973; Zuffa, 1980; Ingersoll et al., 1987; Garzanti et al., 2002b;
Picard and McBride, 2007), but are easily lost in humid climates in the presence of abundant water
and dissolved carbon dioxide (Schnoor and Stumm, 1986; Stumm and Morgan, 1996 p.188ff; Singh
ACCEPTED MANUSCRIPT
21
and France-Lanord, 2002). In arid climates, even gypsum or anhydrite grains are preserved (Fig.
4C; Henares et al., 2014; Garzanti et al., 2016 p.121). There is no objective rule to determine
whether a grain will survive the diverse physico-chemical environments experienced during one or
more sedimentary cycles, and if we are to neglect carbonate rock fragments because they are
potentially soluble, then we should neglect as well even more mechanically or chemically labile
lithic grains such as shale or gypsum. Few sedimentary lithics will be retained.
PT
Petrographic analysis of sand or sandstone samples poses multiple challenges, but they need to be
RI
faced with the conceptual and technological tools at our disposal, and hopefully won with time one
SC
by one. Whenever we fail, problems should be exposed, otherwise they will never be solved. All
types of extrabasinal lithic fragments, independently of their presumed stability, are considered and
NU
assigned to pole L in the proposed classification scheme.
MA
Contrasting compositions of sand derived from diverse orogenic and anorogenic tectonic domains,
and found in major river systems, deserts, and deep-sea fans, provide a solid objective basis to
T
establish a link between petrographic signature and geology of source terranes. Such an approach,
EP
widely pursued since the last century based on both QFR (Krynine, 1948; Folk, 1968; Potter, 1978)
C
and QFL detrital modes (Dickinson and Suczek, 1979), showed that compositional patterns are not
AC
physical and chemical processes during erosion, transportation, and deposition for modern
sediments (environmental bias) and burial as well for ancient sedimentary rocks (diagenetic bias;
Johnsson, 1993; Weltje, 2012; Basu, 2017). An updated overview is provided in this conclusive
section, which focuses on modern sediments where source terranes and their geodynamic,
geomorphological, and climatic settings are known and all factors affecting sediment composition
can be verified and includes unpublished data collected at the Laboratory of Provenance Studies
ACCEPTED MANUSCRIPT
22
(University of Milano-Bicocca) on river sand in North America, northern Europe, northern Asia and
Australia, and on eolian sand from the Kalahari and Sahara sand seas.
Major orogenic and anorogenic sources of igneous detritus include magmatic arcs, ophiolitic
complexes, and continental flood basalts. Main characteristic features are the abundance of feldspar
PT
in plutoniclastic sand, of volcanic lithics in volcaniclastic sand, and of mafic to ultramafic lithics in
RI
ophiolite-derived sand.
SC
Magmatic arcs as paradigmatic sources of igneous detritus have been thoroughly investigated by
W.R. Dickinson and co-workers, who documented how composition changes systematically
NU
through time as the plutonic roots of the arc massif are progressively stripped off of their volcanic
cover (Dickinson, 1985; Marsaglia and Ingersoll, 1992; Ingersoll, 2012). Undissected basaltic to
MA
andesitic volcanic provinces shed quartz-poor sand containing mainly microlitic lithic fragments
and plagioclase, thus plotting in the FL field and less commonly in the LF field (e.g., owing to
ED
hydraulic concentration of crystals; Dickinson, 1970 p.705). Sand derived from more felsic
T
rhyodacitic products contains quartz and largely felsitic volcanic lithics and may straddle the qFL
EP
and fQL fields. Wherever the deep-seated batholithic roots of the arc start to be patchily exposed,
quartz and K-feldspar increase at the expense of volcanic lithics and detrital modes may straddle the
C
qLF, qFL, and fQL fields. At an advanced stage of dissection, detritus will plot in the lQF or lFQ
AC
fields and, finally, where only the plutonic arc basement is exposed, in the QF or fFQ fields
Volcaniclastic detritus shed from continental flood basalts in rift-related settings is not readily
discriminated from detritus derived from undissected magmatic arcs. Large anorogenic igneous
provinces, however, are typically characterized by bimodal products. The Ethiopia-Yemen Traps,
including both basaltic flows and felsic lavas and ignimbrites (Ayalew et al., 2002; Ukstins et al.,
2002), shed sand straddling the L, FL, qFL, and fQL fields (Garzanti et al., 2001, 2015b). The
ACCEPTED MANUSCRIPT
23
Karoo and Etendeka Traps of southern Africa, the latter including quartz latites at higher
stratigraphic levels (Ewart et al., 2004), supply sand plotting in the L and subordinately FL fields
(Garzanti et al., 2014). The potassic Virunga volcanoes, including silica-saturated latites and
trachytes (Rogers et al., 1998), generate sand containing very little or no quartz and plotting in the
FL and subordinately L fields (Garzanti et al., 2013a). Sand produced in basaltic oceanic islands,
including Iceland, Cape Verde, Tahiti, and Hawaii, also plot in the FL and L fields (Marsaglia,
PT
1993; Dinis et al., 2019).
RI
Because the only quartz that oceanic lithosphere contains is hosted in plagiogranite bodies at the top
SC
of magma chambers (Dilek and Furnes, 2011), ophiolite-derived sand consists almost entirely of
plagioclase and mafic to ultramafic lithic fragments. The ideal petrographic trend in case of
NU
progressive dissection of oceanic lithosphere includes volcaniclastic detritus with dominant
lathwork basaltic to vitric and locally boninitic lithics derived from pillow lavas and sheeted-dike
MA
complexes, followed by plutoniclastic detritus derived from underlying plagiogranite and gabbro,
and finally by ultramafic detritus dominated by cellular serpentinite lithics derived from obducted
ED
mantle rocks (Garzanti et al., 2000, 2002a). In the QFL triangle, such a trend would describe a
counter-clockwise loop starting from the FL field for basalticlastic sand, moving towards the qLF
T
EP
or even F field for theoretically pure gabbro sand, and finally back to the L field for ultramafic sand
Major orogenic and anorogenic sources of sedimentary and metamorphic detritus are fold-thrust
belts, subduction complexes, and crustal sections exposed on rift shoulders or within continental
blocks. Sand composition principally depends on the tectono-stratigraphic level exposed to erosion
in the source area. This was first perceived by Krynine (1948), who envisaged the continental crust
rocks generating lithic-rich “greywacke,” and eventually deeper-seated plutonic rocks generating
Beside quartz, however, sedimentary rocks can shed a wide spectrum of lithic fragments, and
detrital modes of sedimentaclastic sand may consequently plot all along the QL leg of the QFL
triangle (e.g., Graham et al., 1976). Depending on their mudrock/sandstone ratio, turbiditic
successions accreted within subduction complexes shed sand plotting in the L and QL fields (e.g.,
PT
Garzanti et al., 1998; Di Giulio et al., 2003), in the QL and LQ fields, or even in the Q and qQ
RI
fields (Garzanti et al., 2013b; Limonta et al., 2015). Wherever parent turbiditic sandstones within
SC
accretionary prisms include common feldspar, daughter sand may plot in the fQL, fLQ, or even in
the lFQ field (Cavazza et al., 1993; Garzanti et al., 2002b; Fontana et al., 2003).
NU
At another extreme, sedimentaclastic detritus genenerated in both orogenic and anorogenic settings
field. Sand consisting exclusively of limestone and dolostone grains (Fig. 4 A,B) is found in
tropical to middle-latitude regions, including the European Alps, the Apennines, and the Middle
ED
East from the Levant rifted margin to the Zagros Mountains, where arid or semiarid climate
favoured carbonate rather than siliciclastic production along Neotethyan shores through most of the
T
EP
Mesozoic. In the same regions, sedimentaclastic sand characterized or even dominated by chert
grains is derived from pelagic strata deposited originally along distal passive margins of southern
C
Neotethys (e.g., Toscana-Umbria domain in the Apennines, Mamonia complex of south Cyprus,
AC
progressive increase in metamorphic grade. During unroofing of upper crustal levels, quartz
increases from slate sand derived from anchimetamorphic and epimetamorphic metapelites, to
phyllite sand derived from lower-greenschist or blueschist facies schists and calcschists, all plotting
in the QL field. Schist sand derived from upper greenschist-facies schists, paragneisses and marbles
may reach into the fLQ field (Garzanti et al., 2010a). During unroofing of deeper crustal levels,
ACCEPTED MANUSCRIPT
25
feldspar increases in gneiss sand derived from amphibolite-facies granitoid gneisses and plotting in
the lFQ field, in kinzigite sand derived from upper-amphibolite-facies metasediments and straddling
the lFQ and lQF fields, and in stronalite sand derived from lower-crustal granulite-facies
metasediments associated with metagabbros and straddling the lQF and qLF fields (Garzanti et al.,
2006).
As a peculiar case of great provenance value, detritus from mantle rocks that have undergone
PT
eclogite-facies metamorphism during subduction may consist entirely of strongly foliated
RI
antigorite-serpentineschist grains plotting in the L field (Fig. 4P; Garzanti et al., 1998, 2004).
SC
8.3. Large rivers
NU
Transcontinental rivers drain diverse geological domains and are thus typically characterized by
mixed sediment compositions (third-order sampling scale of Ingersoll et al., 1993). Quartz, which is
MA
both widespread in source rocks and durable, commonly predominates. Lithic grains vary in type
and relative abundance, whereas feldspar is generally subordinate but common in case of
ED
Large rivers draining subequatorial cratonic regions in Africa or South America may carry sand
EP
consisting almost exclusively of monocrystalline quartz and thus plotting in the qQ field (Congo,
Okavango, White Nile, Paranà, Uruguay), or containing a few feldspars and plotting in the Q
C
(Niger) or even FQ fields (Zambezi). Sand of the Orinoco River, draining also the Andean retroarc
AC
basin, includes a few sedimentary and low-rank metasedimentary lithics (Q; Limonta et al., 2015).
Feldspatho-quartzose river sand is found at tropical (e.g., Limpopo sand, including volcanic lithics
from anorogenic Karoo basalts; fFQ), middle (Mississippi sand; FQ), and high latitudes (e.g.,
Large rivers sourced in orogenic belts in Asia, Europe, or South America carry diverse sedimentary
and metamorphic lithics, and their sand may plot in the fLQ field (Rhein, Po, Danube, Kuban,
Amazon), in the lFQ field (Brahmaputra, Yellow River, Red River, Mekong, Salween; Borges et
ACCEPTED MANUSCRIPT
26
al., 2008; Garzanti et al., 2010b; Nie et al., 2015), or straddle the fLQ and lFQ fields (Indus, Ganga,
Irrawaddy,Yang Tze; Garzanti et al., 2005, 2010b, 2016; Vezzoli et al., 2016). Volcanic lithics
derived from andesitic volcanoes in the Andes and Greater Caucasus are common in sand of the
Amazon and Kuban Rivers, whereas volcanic lithics from the anorogenic Emeishan basalts are
subordinate in Yang Tze sand. Sand of several major rivers in North America and Russia also plot
in the fLQ (Slave, Liard, Colorado) or lFQ fields (St. Lawrence, Lena, Ural), together with bedload
PT
sand of transcontinental rivers such as the Nile and the Orange, which drain continental blocks
RI
including vast anorogenic basaltic provinces and thus carry notable amounts of largely mafic
SC
volcanic lithics. Nile suspended load is instead almost purely volcaniclastic and straddles the qFL
and fQL fields, as sand of the Columbia River that drains both orogenic and anorogenic volcanic
NU
rocks (Whetten et al., 1969).
Quartz-poor river sand plotting in the fQL field is typical of largely undissected orogenic domains
MA
including volcanic regions, such as the Caucasus and the Anatolian-Iranian Plateau (Terek, Rioni,
Kura, Tigris, Euphrates; Vezzoli et al., 2014; Garzanti et al., 2016). European rivers sourced in the
ED
Alpine, Apenninic, or Pyrenean thrust belts and largely draining sedimentary strata may carry
None of the considered rivers carries feldspar-dominated sand, but feldspar-poor examples are also
rare. They include the sedimentaclastic sand of the Murray-Darling River in Australia (LQ) and of
C
the Peel River in Canada (QL), as well as the almost purely carbonaticlastic sand of the Karun
AC
River draining the Zagros Mountains in Iran and of several rivers draining the eastern European
Alps (e.g., Piave, Tagliamento; Garzanti et al., 2006b, 2016; Picard and McBride, 2007).
The composition of deep-sea turbidites chiefly reflects that of their main fluvial feeder system
(Dickinson, 1988; Zuffa et al., 2000). Sand in the huge Indus and Bengal-Nicobar Fans, supplied
from the Himalayas throughout the Neogene, plots in the lFQ field (Ingersoll and Suczek, 1979;
ACCEPTED MANUSCRIPT
27
Suczek and Ingersoll, 1985; Pickering et al., 2018; Andò et al., 2019). Litho-quartzose
metamorphiclastic turbidites characterize the Hellenic Trench (Bartolini et al., 1975), and quartzose
derived turbiditic sandstones exposed on Barbados island (Velbel, 1985; Limonta et al., 2015).
Trench, forearc, and back-arc turbidites all around the Pacific Ocean are chiefly fed from erosion of
magmatic arcs in various stages of dissection (Fig. 7; Dickinson, 1982; Thornburg and Kulm,
PT
1987). Deep-sea sand derived from undissected island arcs or continental arcs mostly consists of
RI
volcanic lithics and plagioclase, and thus plots in the FL field or occasionally in the LF field.
SC
Increasing degrees of dissection are indicated by progressive increase in quartz, K-feldspar,
sedimentary or metamorphic lithics, until only quartz, feldspars, and lithic grains from metamorphic
NU
wallrocks are found where granitoid batholiths are stripped off of their volcanic and sedimentary
cover strata (Marsaglia and Ingersoll, 1992). Sand thus plots in the qFL field or in the adjacent fQL
MA
and qLF fields in the transitional stage (e.g., Yerino and Maynard, 1984; De Rosa et al., 1986;
Marsaglia et al., 1995; Heberer et al., 2010), and straddle the lQF and lFQ fields or even the QF and
ED
fFQ fields when the arc basement becomes the dominant or exclusive source (Bachman and
Leggett, 1982).
T
EP
Quaternary sand in deep-sea fans fed by transcontinental rivers draining anorogenic continental
domains largely consists of quartz and feldspar. Congo Fan sand is quartzose, whereas Nile Cone
C
sand plots in the qFQ field and contains volcanic lithics derived from the Ethiopian continental
AC
8.5. Deserts
Dune fields occupy wide areas in arid tropical and subtropical regions. Quartz is commonly
dominant because of its mechanical and chemical durability (Muhs, 2004), and eolian sand plots
mostly in the qQ field where it is largely recycled from older quartzose sandstone, as in the
Kalahari Desert of southern Africa or in the Nafud and Dahna Deserts of Arabia (Fig. 8; Garzanti et
ACCEPTED MANUSCRIPT
28
al., 2013c, 2014). For the same reason, quartz is dominant across the Sahara, although dunes in the
Western Desert of Egypt and in Sinai may contain feldspar and mainly carbonate rock fragments
closer to exposed Mesozoic and Cenozoic carbonates, thus plotting in the qFQ or lFQ fields. Sand
accumulated in coastal deserts along the hyperarid Atlantic coast of southwestern Africa, all largely
fed from the Orange River, includes significant basaltic detritus and plots in the FQ or lFQ fields.
Sand seas of central Asia, such as the Taklamakan and Karakum, largely fed by the Yarkhand and
PT
Amu Darya fluvial systems, commonly include carbonate and other sedimentary or
RI
metasedimentary lithics and may plot in the fQL, fLQ, or lFQ fields (Rittner et al., 2016; Garzanti
SC
et al., 2019).
NU
9. CONCLUSIONS
MA
“The perfect classification of sandstones does not now and never will exist” (Folk, 1980, p.126).
Indeed, the information contained in a sand or sandstone sample is so rich that it could never be
ED
compressed in a short classificatory label. Nature is never trivially repetitive, and the interplay of
physical and chemical processes that generate sand produces an endless suite of products that
T
cannot be captured in a finite set of standard cages. Most classifications proposed in the past have
EP
revealed the ambiguities inevitably implied in the hopeless attempt to encapsulate in one or two
words and worse when referring to vaguely defined archetypes such as greywacke or arkose a
C
AC
range of different properties, including mineralogy and texture as well as hydraulic behaviour, and
the presumed durability of detrital grains to mechanical or chemical attack during one or more
sedimentary cycles. This illusion has generated a conceptual tangle that needs to be unraveled.
This paper reviews the somewhat intricate problems involved in the classification of sand and
sandstone, and prescribes a series of norms apt to communicate efficiently the results of
petrographic observations in a scientific article or technical report. For quantitative analysis, the
Gazzi-Dickinson method is recommended in order to get rid of the geometrical effects related to
ACCEPTED MANUSCRIPT
29
breakage of coarse-grained rock fragments into their monomineralic constituents (Ingersoll et al.,
1984; Zuffa, 1985), and thus isolate the component of grain-size-dependent intrasample and
intersample compositional variability associated with provenance and hydraulic sorting in the
sheet (e.g., Fig. 5) that allows collection of data on the complete spectrum of rock fragments in the
sample and to recalculate petrographic parameters according to both Gazzi-Dickinson QFL and
PT
traditional QFR methods. All quartz, both monocrystalline and polycrystalline, should be included
RI
in pole Q, and all sedimentary lithics, including chert, extrabasinal carbonate and evaporite, should
SC
be included in pole L (Zuffa, 1980).
The essential information thus obtained can be condensed in a composite adjective based on the
NU
relative abundance of quartz, feldspar, and lithic fragments. According to standard use (Crook,
1960; Dickinson, 1970), the less abundant component goes first, the more abundant last. For lithic-
MA
rich sand and sandstone, information on the prevailing rock fragment type can be communicated by
an additional adjective (Ingersoll, 1983). For lithic-poor sand and sandstone, further formal
ED
subdivisions are proposed, thus reaching a total of 18 compositional fields overall. This scheme
proves to be sufficiently complete to formulate a clear description of any sand or sandstone sample
T
EP
intended to be short and exhaustive, precise and flexible at the same time, as documented with
examples of modern sand in major rivers, eolian dune fields, and deep-sea fans.
C
AC
ACKNOWLEDGMENTS
This article is dedicated to Bob Folk and Bill Dickinson, who taught me far more than the riddles of
sandstone petrography. Ray Ingersoll and Gert Jan Weltje kindly provided masterly reviews,
advice, and extremely useful constructive critical comments. Unpublished petrographic data used in
this study were obtained by Giovanni Vezzoli, Alberto Resentini, and Matteo Sala from samples
ACCEPTED MANUSCRIPT
30
kindly provided by A.Haedke and H.Wittman (world rivers), A. Stone (Kalahari), Y. Najman, Laura
differences in nomenclature and definition of end-members and compositional fields. Chert is either
grouped with quartz in Krynine’s, Hubert’s, early Folk’s, and McBride’s QFR plots to emphasize
PT
durability, but with rock fragments in van Andel’s and late Folk’s QFR plots to emphasize
provenance. In the genetic Dickinson’s scheme, based instead on the Gazzi-Dickinson method and
RI
on the inferred correspondence between detrital modes and geodynamic setting, chert is grouped
SC
with quartz in the QtFL plot and with lithic fragments in the QmFLt plot, whereas carbonate rock
NU
fragments are excluded.
MA
Figure 2. Traditional sandstone classifications based on both petrographic composition and texture.
In Pettijohn’s, Gilbert’s, Packham’s, and Crook’s schemes, two different triangular diagrams are
ED
diagrams were also devised to take into account what Pettijohn (1954 p.360, 363) believed to be the
EP
three factors “of greatest genetic importance in the classification of sandstones”, i.e: “provenance,
C
maturity, and fluidity of the depositing medium”. Adding texture as a fourth component of the
AC
sacred QFR triad, however, involves nomenclatural, graphical, and conceptual confusion (e.g.,
turbidites are not necessarily matrix-rich, pseudomatrix being commonly generated by post-
depositional alteration of labile rock fragments; Okada, 1966; Whetten, 1969; Dickinson, 1970).
Figure 3. The proposed classification of sand and sandstone. A) Straightforward subdivision into
quartzose, feldspathic, and lithic compositions by considering the most abundant QFL component
only. B) In the scheme of Weltje (2006), 6 fields are obtained by considering the relative abundance
ACCEPTED MANUSCRIPT
31
of the two most abundant QFL components. C) In the expanded scheme of Garzanti (2016), 15
fields are obtained by considering all three QFL components provided they exceed 10%QFL. D)
The refined classification of feldspatho-quartzose and quartzose sand and sandstone (Garzanti et al.,
2018a) leads to 18 fields overall. Q = quartzose (qQ = pure quartzose); F = feldspathic; L = lithic;
PT
feldspatho-quartzose; lQF = litho-quartzo-feldspathic; qLF = quartzo-litho-feldspathic; qFL =
RI
quartzo-feldspatho- lithic; fQL = feldspatho-quartzo- lithic; fLQ = feldspatho-litho-quartzose.
SC
Figure 4. Rock fragments: the goldmine of provenance information. Sedimentary grains: A, B)
NU
carbonaticlastic sand (Wadi Bih, northern Oman); C) gypsum sand (Azraq dune field, Jordan); D)
chert sand (Romandato creek, southern Italy); E) clast reworked from Oligocene turbidites
MA
(Galathea River, Great Nicobar Island). Metasedimentary grains: F) slate sand (Dazhu River,
Taiwan); G) calcschist from the Tauern Window (Fuschbach River, Austria); H) fibrolitic-
ED
sillimanite schist from the Greater Himalaya (Dordi Khola, Nepal). Volcanic grains derived from:
I) Deccan Trap basalts (Tapti River, India), J) andesites (Rio Grande, Argentina), and K) rhyolites
T
EP
(Lipari Island, Italy). L) Metavolcanic grains from anchimetamorphic Permian rhyolites (Southern
Alps, Italy). Metabasite grains: M) epidote greenschist (Laba River, NW Greater Caucasus) and
C
cellular serpentinite from obducted ophiolites (Wadi Ham, northern Oman); P) foliated
serpentineschist from subducted metaophiolites (Voltri beach, Italy). All photos taken with crossed
Figure 5. Point-counting sheet used routinely for petrographic analysis of modern sand at the
Figure 6. Sand composition and classification in the world’s largest rivers (mostly own published
and unpublished data). Quartz and sedimentary lithics are generally predominant and feldspar
subordinate (Q/F never < 1) , indicating that in most fluvial systems a large part of the sediment
PT
load is recycled from sedimentary strata. Pure quartzose sand characterizes rivers draining
RI
continental blocks at subequatorial latitudes, where both weathering and recycling are extensive
SC
(White Nile, Congo, Okavango). Lithic-rich sand characterizes rivers draining extensive
sedimentary and volcanic covers in undissected orogenic domains (Euphrates, Tigris, Terek, Kura,
NU
Rhône, Ebro, Tevere). Volcanic lithics are predominant in rivers draining orogenic andesitic arcs
(Amazon), anorogenic basaltic fields (Nile, Orange, Limpopo), or both (Columbia; Whetten et al.,
MA
1969). Metamorphic lithics are predominant in rivers draining either very active orogens (e.g.,
Brahmaputra) or quiescent continental shields (e.g., Glomma). Smaller symbols refer to suspended
ED
load and exemplify two opposite cases: composition is quite similar to bedload in the Irrawaddy
River, but markedly different in the Nile River. Q = quartz; F = feldspar; L = lithics (Lm =
T
EP
Figure 7. Sand composition and classification in deep-sea fans. Trench, forearc and back-arc
AC
turbidites all around the Pacific Ocean are chiefly fed from erosion of island arcs in the west and
continental arcs in the east. Relative abundance of volcanic lithics, feldspar, and quartz depends on
character of magmatism and degree of dissection of source terranes (Dickinson, 1985). Remnant-
ocean turbidites shed from the Himalayan orogen are rich in quartz and metamorphic or
sedimentary lithics. Deep-sea fans supplied from transcontinental rivers draining continental blocks
are quartz-rich (e.g., Congo Fan) but may include feldspar and volcanic lithics derived from
continental-flood basalts (e.g., Nile Cone). Each point is the mean sand composition from the upper
ACCEPTED MANUSCRIPT
33
part of one deep-sea core. Fields for magmatic-arc provenance are based on data compilations in
Marsaglia and Ingersoll (1992), Ingersoll and Eastmond (2007), and Garzanti et al. (2018c). Data
sources: BL = Bachman and Leggett (1982); Bl = Baltuck et al. (1985); Br = Bartolini et al. (1975);
D = De Rosa et al. (1986); GI = Gergen and Ingersoll (1986); G = Garzanti et al. (2018b) and own
data; IS = Ingersoll and Suczek (1979); M = Marsaglia (2004); Marsaglia et al. (1992, 1995); PH =
Prasad and Hesse (1982); PI = Packer and Ingersoll (1986); S = Stewart (1977, 1978); SI = Suczek
PT
and Ingersoll (1985); YM = Yerino and Maynard (1984). Q = quartz; F = feldspar; L = lithics (Lm
RI
= metamorphic; Lv = volcanic; Ls = sedimentary).
SC
Figure 8. Sand composition and classification in the world’s largest deserts. Quartz is predominant
NU
in the Sahara, Kalahari, and Great Nafud sand seas nourished by extensive recycling of older
quartz-rich sandstones. Limestone lithics may be common in the Rub’ al Khali and other Arabian
MA
deserts and also occur in dunes of Egypt and Sinai. Basaltic lithics and plagioclase characterize
dune fields along the hyperarid coast of Namibia and Angola mostly fed by the Orange River.
ED
Sedimentary and metasedimentary lithics are common in central Asia dune fields, largely supplied
by major rivers draining orogenic belts (Amu Darya, Yarkhand, Yellow River). Q = quartz; F =
T
EP
Table 1. Criteria used in the classification of sandstone through the last 70 years.
AC
- te M
PT
quartz arkosic lithic
Gilber 19 X X - X pole Q pole pole R --- ?
arenit arenit areni
t 54 - Q
- e/ e/ te /
w ack w acke w ack
RI
e e
Bokm 19 X X - X orthoquar ark lithic sdst. pole Q pole ? ? pol
SC
an 55 - tzite os Q e
- e M
Packh 19 X X - X quartzose arkose / labile pole Q pole ? ? pol
am 54 - sdst./ Q e
labile sdst. /
- subgrayw M
grayw ac labile
NU
acke ke grayw ack
e
quartzo feldspath lithic /
Crook 19 X (X) - X pole Q pole --- ? pol
se ic / labile labile
MA
60 - R e
arenit arenite arenit
- M
e e
Huber 19 X X X X orthoquar ark grayw ack pole Q pole --- po ---
t 60 tzite os e Q le
e R
ED
e/ / w acke te /
w ack w ack
e e
Gazzi 19 X X - (X) (protoquar arc (arenite Pole Q pole --- --- ---
C
66 - zite) os litica) Q
- e
AC
Folk 19 X --- ( (X) quartzare ark litharenite Pole Q pole pole R --- ---
68- X nite os R
80 ) e
Dickin 19 X X - --- quartzo feldspathic lithic Pole Q pole --- --- ---
son 70 - Q
se sdst. sdst.
-
sandsto (arkose (grayw a
nes ) cke)
quartzo feldspathic
Dickin 19 X X X --- lithic Poles Qt Pol --- --- ---
se sdst.,
son 85 sandston & Lt es
sandsto arkose e Qt
nes & Lt
Zuffa 19 X --- - --- quartzare ark litharenite Pole Q pole Pole L --- ---
80- - nite os L
85 - e
This 20 X --- - --- quartzo feldspath lithic Pole Q pole Pole L --- ---
one 19 - sand/ston L
se ic
- e
sand/sto sand/sto
ne ne
ACCEPTED MANUSCRIPT
35
REFERENCES
Ager, D.V., 1981. The Nature of the Stratigraphical Record. Halsted Press, New York, 122 p.
Andò, S., Aharonovich, S., Hahn, A., George, S.C., Clift, P.D., Garzanti, E., 2019. Integrating heavy-mineral,
geochemical, and biomarker analyses of Plio-Pleistocene sandy and silty turbidites: a novel approach for
provenance studies (Indus Fan, IODP Expedition 355). Geological Magazine, in press.
PT
Ayalew, D., Barbey, P., Marty, B., Reisberg, L., Yirgu, G., Pik, R., 2002. Source, genesis, and timing of giant
ignimbrite deposits associated with Ethiopian continental flood basalts. Geochimica et Cosmochimica
RI
Acta, 66, 1429-1448.
SC
Bachman, S.B., Leggett, J.K., 1982. Petrology of Middle America Trench and trench slope sands, Guerrero
Margin, Mexico. Initial reports of the Deep Sea Drilling Project, 66, 429-436.
NU
Baltuck, M., von Huene, R., Arnott, R.J., 1985. Sedimentology of the western continental slope of Central
America. Initial reports of the Deep Sea Drilling Project, 84, 921-937.
MA
Bartolini, C., Malesani, P.G., Manetti, P., Wezel, F.C., 1975. Sedimentology and petrology of Quaternary
sediments from the Hellenic trench, Mediterranean ridge and the Nile cone from D.S.D.P., Leg 13, cores.
Sedimentology, 22, 205–236.
ED
Basu, A., 2017. Evolution of siliciclastic provenance inquiries: A critical appraisal. In: Mazumder, R. (Ed.),
Sediment Provenance, Influences on Compositional Change from Source to Sink. Elsevier, ch.2, 5-23,
T
http://dx.doi.org/10.1016/B978-0-12-803386-9.00002-2.
EP
Blatt, H., Middleton, G., Murray, R., 1972. Origin of Sedimentary Rocks, 2nd Edition. Prentice-Hall,
Englewood, New York, 782 p.
C
Bokman, J., 1955, Sandstone classification: Relation to composition and texture. Journal of Sedimentary
AC
Boswell, P.G.H., 1960. The term graywacke. Journal of Sedimentary Petrology, 30, 154-157.
Brongniart, A., 1826. L’ arkose, caractères minéralogiques et histoire géognostique de cette roche. Annales des
sciences naturelles, 8, 113-163.
Cavazza, W., Zuffa, G.G., Camporesi, C., Ferretti, C., 1993. Sedimentary recycling in a temperate climate
drainage basin (Senio River, north-central Italy): composition of source rock, soil profiles, and fluvial
deposits. In: Johnsson, M.J., and Basu, A. (Eds.), Processes Controlling the Composition of Clastic
Sediments. Geological Society of America, Special Paper, 284, 247-262.
Chen, P.Y., 1968. A modification of sandstone classification. Journal of Sedimentary Petrology, 38, 54-60.
ACCEPTED MANUSCRIPT
36
Cleary, W.J., Conolly, J.R., 1972. Embayed quartz grains in soils and their significance. Journal of Sedimentary
Petrology, 42, 899-904.
Critelli, S., Ingersoll, R.V., 1995. Interpretation of neovolcanic versus palaeovolcanic sand grains: an example
from Miocene deep‐marine sandstone of the Topanga Group (Southern California). Sedimentology, 42, 83-
804.
Crook, K.A.W., 1960. Classification of arenites. American Journal of Science, 258, 419-428.
Crook, K.A.W., 1968. Weathering and roundness of quartz sand grains. Sedimentology, 11, 171-182.
PT
Crook, K.A.W., 1974. Lithogenesis and geotectonics: The significance of compositional variation in flysch
arenites (graywackes). In: Dott, R.H., Shaver, R.H. (Eds.), Modern and Ancient Geosynclinal
RI
Sedimentation. Society of Economic Paleontologists and Mineralogists, Special Publication 19, 304-310.
SC
Dapples, E.C., Krumbein, W.C., Sloss, L.L., 1953. Petrographic and lithologic attributes of sandstones. The
Journal of Geology, 61, 291-317.
NU
De Rosa, R., Zuffa, G.G., Taira, A., Leggett, J.K., 1986. Petrography of trench sands from the Nankai Trough,
southwest Japan: Implications for long-distance turbidite transportation. Geological Magazine, 123, 477-
MA
486.
Dickinson, W.R., 1970. Interpreting detrital modes of graywacke and arkose. Journal of Sedimentary Petrology,
40, 695-707.
ED
Dickinson, W.R., 1982, Compositions of sandstones in circum-Pacific subduction complexes and forearc basins.
American Association of Petroleum Geologists Bulletin, 66, 121-137.
T
EP
Dickinson, W.R., 1985. Interpreting provenance relations from detrital modes of sandstones. In: Zuffa, G.G.
(Ed.), Provenance of arenites. Reidel, Dordrecht, NATO ASI Series 148, 333-361.
C
Dickinson, W.R., 1988. Provenance and sediment dispersal in relation to paleotectonics and paleogeography of
sedimentary basins. In: Kleinspehn, K.L., Paola, C. (Eds.), New perspectives in basin analysis. Springer,
AC
Berlin, 3-25.
Dickinson, W.R., 2003. The place and power of myth in geoscience: an associate editor's perspective. American
Journal of Science, 303, 856-864.
Dickinson, W.R., Suczek, C.A., 1979. Plate tectonics and sandstone composition. American Association of
Petroleum Geologists Bulletin, 63, 2164-2172.
Di Giulio, A., Ceriani, A., Ghia, E., Zucca, F., 2003. Composition of modern stream sands derived from
sedimentary source rocks in a temperate climate (Northern Apennines, Italy). Sedimentary Geology, 158,
145-161.
ACCEPTED MANUSCRIPT
37
Dilek, Y., Furnes, H., 2011. Ophiolite genesis and global tectonics: Geochemical and tectonic fingerprinting of
ancient oceanic lithosphere. Geological Society of America Bulletin, 123, 387-411.
Dinis, P.A., Pinto, M.C., Rocha, F.T., Garzanti, E., 2019. Detrital record of the denudation of volcanic islands
under sub-tropical climate (Cape Verde). Chemie der Erde, in press.
Doglioni, C., Panza, G.F., 2015. Polarized plate tectonics. Advances in Geophysics, 56, 1-167.
Dorsey, R.J., 1988. Provenance evolution and unroofing history of amodern arc–continent collision: Evidence
from petrography of Plio–Pleistocene sandstones, eastern Taiwan. Journal of Sedimentary Petrology, 58,
PT
208-218.
Dott, R.H., 1964. Wacke, graywacke and matrix - what approach to immature sandstone classification? Journal
RI
of Sedimentary Petrology, 34, 625-632.
Ewart, A., Marsh, J.S., Milner, S.,C., Duncan, A.R., Kamber, B.S., Armstrong, R.A., 2004. Petrology and
SC
geochemistry of Early Cretaceous bimodal continental flood volcanism of the NW Etendeka, Namibia.
Journal of Petrology, 45, 59-138.
NU
Folk, R.L., 1954. The distinction between grain size and mineral composition in sedimentary rock
nomenclature. The Journal of Geology, 62, 344-359.
MA
Folk, R.L., 1968 and 1980, Petrology of Sedimentary Rocks. Austin (USA), Hemphill Publishing Co., 182 p.
ED
Fontana, D., Zuffa, G.G., Garzanti, E., 1989. The interaction of eustacy and tectonism from provenance studies
of the Eocene Hecho Group Turbidite Complex (South Central Pyrenees, Spain). Basin Research, 2, 223-
237.
T
EP
Fontana, D., Parea, G.C., Bertacchini, M., Bessi, P., 2003. Sand production by chemical and mechanical
weathering of well lithified siliciclastic turbidites of the Northern Apennines (Italy). In: Valloni, R., Basu,
A. (Eds.), Quantitative Provenance Studies in Italy. Memorie Descrittive della Carta Geologica d' Italia, 61,
C
51-60.
AC
Füchtbauer, H., 1959. Zur Nomenklatur der Sedimentgesteine. Erdöl und Kohle, 12, 605-613.
Galloway, W.E., 1974. Deposition and diagenetic alteration of sandstone in northeast Pacific arc-related basins:
Implications for graywacke genesis. Geological Society of America Bulletin, 85, 379-390.
Garzanti, E., 1991. Non-carbonate intrabasinal grains in arenites: Their recognition, significance and
relationship to eustatic cycles and tectonic setting. Journal of Sedimentary Petrology, 61, 959-975.
Garzanti, E., 2016. From static to dynamic provenance analysis—Sedimentary petrology upgraded. Sedimentary
Geology, 336, 3-13.
Garzanti E., 2017. The maturity myth in sedimentology and provenance analysis. Journal of Sedimentary
Research, 87, 353-365
ACCEPTED MANUSCRIPT
38
Garzanti, E., Vezzoli, G., 2003. A classification of metamorphic grains in sands based on their composition and
grade. Journal of Sedimentary Research 73, 830-837.
Garzanti, E., Scutellà, M., Vidimari, C., 1998. Provenance from ophiolites and oceanic allochtons: modern
beach and river sands from Liguria and the Northern Apennines (Italy). Ofioliti, 23/2, 65-82.
Garzanti, E., Andò, S., Scutellà, M., 2000. Actualistic ophiolite provenance: the Cyprus Case. The Journal of
Geology, 108, 199-218.
Garzanti, E., Vezzoli, G., Andò, S., Castiglioni, G., 2001. Petrology of rifted-margin sand (Red Sea and Gulf of
PT
Aden, Yemen). The Journal of Geology, 109, 277-297.
Garzanti, E., Vezzoli, G., Andò, S., 2002a. Modern sand from obducted ophiolite belts (Oman, U.A.E.). The
RI
Journal of Geology, 110, 371-391.
Garzanti, E., Canclini, S., Moretti Foggia, F., Petrella, N., 2002b. Unraveling magmatic and orogenic
SC
provenances in modern sands: The back-arc side of the Apennine thrust-belt (Italy). Journal of Sedimentary
Research, 72, 2-17.
NU
Garzanti, E., Andò, S., Vezzoli, G., Dell’Era, D., 2003. From rifted margins to foreland basins: Investigating
provenance and sediment dispersal across desert Arabia (Oman, UAE). Journal of Sedimentary Research,
MA
73, 572-588.
Garzanti, E., Vezzoli, G., Lombardo, B., Andò, S., Mauri, E., Monguzzi, S., Russo, M., 2004. Collision orogen
provenance (Western and Central Alps): detrital signatures and unroofing trends. The Journal of Geology,
ED
112, 145-164.
Garzanti, E., Andò, S., Vezzoli, G., Ali Abdel Megid, A., El Kammar, A., 2006a. Petrology of Nile River sands
T
(Ethiopia and Sudan): Sediment budgets and erosion patterns. Earth and Planetary Science Letters, 252,
EP
327-341.
Garzanti, E., Andò, S., Vezzoli, G., 2006b. The continental crust as a source of sand (Southern Alps cross-
C
Garzanti E., Andò S., Vezzoli G., 2008. Settling equivalence of detrital minerals and grain-size dependence of
sediment composition. Earth and Planetary Science Letters, 273, 138-151.
Garzanti, E., Andò, S., Vezzoli, G., 2009. Grain-size dependence of sediment composition and environmental
bias in provenance studies. Earth and Planetary Science Letters, 277, 422–432.
Garzanti, E., Resentini, A., Vezzoli, G., Andò, S., Malusà, M.G., Padoan, M., Paparella, P., 2010a. Detrital
fingerprints of fossil continental-subduction Zones (axial belt provenance, European Alps). The Journal of
Geology, 118, 341-362.
ACCEPTED MANUSCRIPT
39
Garzanti, E., Andò, S., France-Lanord, C., Vezzoli, G., Galy, V., Najman, Y., 2010b. Mineralogical and
chemical variability of fluvial sediments. 1. Bedload sand (Ganga–Brahmaputra, Bangladesh). Earth and
Planetary Science Letters, 299, 368-381.
Garzanti, E., Andò, S., Vezzoli, G., Lustrino, M., Boni, M., Vermeesch, P., 2012. Petrology of the Namib sand
sea: Long-distance transport and compositional variability in the wind-displaced Orange Delta. Earth-
Science Reviews, 112, 173-189.
Garzanti, E., Padoan, M., Andò, S., Resentini, A., Vezzoli, G., Lustrino, M., 2013a. Weathering and relative
durability of detrital minerals in equatorial climate: Sand petrology and geochemistry in the East African
PT
Rift. The Journal of Geology, 121, 547-580.
RI
Garzanti, E., Limonta, M., Resentini, A., Bandopadhyay, P.C., Najman, Y., Andò, S., Vezzoli, G., 2013b.
Sediment recycling at convergent plate margins (Indo-Burman Ranges and Andaman-Nicobar Ridge).
SC
Earth-Science Reviews, 123, 113-132.
Garzanti, E., Vermeesch, P., Andó, S., Vezzoli, G., Valagussa, M., Allen, K., Khadi, K.A., Al-Juboury, I.A.,
NU
2013c. Provenance and recycling of Arabian desert sand. Earth-Science Reviews, 120, 1-19.
Garzanti, E., Vermeesch, P., Padoan, M., Resentini, A., Vezzoli, G., Andò, S., 2014. Provenance of passive-
MA
Garzanti, E., Resentini, A., Andò, S., Vezzoli, G., Vermeesch, P., 2015a. Physical controls on sand composition
and relative durability of detrital minerals during long-distance littoral and eolian transport (coastal
ED
Garzanti, E., Andò, S., Padoan, M., Vezzoli, G., El Kammar, A., 2015b. The modern Nile sediment system:
T
Garzanti, E., Al-Juboury, A.I., Zoleikhaei, Y., Vermeesch, P., Jotheri, J., Akkoca, D.B., Allen, M., Andò, S.,
Limonta, M., Padoan, M., Resentini, A., Rittner, M., Vezzoli, G., 2016. The Euphrates-Tigris-Karun river
C
system: Provenance, recycling and dispersal of quartz-poor foreland-basin sediments in arid climate. Earth-
Science Reviews, 162, 107-128.
AC
Garzanti, E., Radeff, G., Malusà, M., 2017a. Slab breakoff: A critical appraisal of a geological theory as applied
in space and time. Earth-Science Reviews, 177, 303-319.
Garzanti E., Vermeesch, P., Al-Ramadan, K.A., Andò, S., Limonta, M., Rittner, M., Vezzoli, G., 201b7. Tracing
transcontinental sand transport: From Anatolia-Zagros to the Rub' al Khali Sand Sea. Journal of
Sedimentary Research, 87, 1196-1213.
Garzanti, E., Dinis, P., Vermeesch, P., Andò, S., Hahn, A., Huvi, J., Limonta, M., Padoan, M., Resentini, A.,
Rittner, M., Vezzoli, G., 2018a. Dynamic uplift, recycling, and climate control on the petrology of passive-
margin sand (Angola). Sedimentary Geology, 375, 86-104.
ACCEPTED MANUSCRIPT
40
Garzanti, E., Andò, S., Limonta, M., Fielding, L., Najman, Y., 2018b. Diagenetic control on mineralogical
suites in sand, silt, and mud (Cenozoic Nile Delta): Implications for provenance reconstructions. Earth-
Science Reviews, 185, 122-139.
Garzanti, E., Limonta, M., Vezzoli, G., An, W., Wang, J.G., Hu, X.M., 2018c. Petrology and multimineral
fingerprinting of modern sand generated from a dissected magmatic arc (Lhasa River, Tibet). In: Ingersoll,
R.V., Lawton, T.F., Graham, S.A. (Eds.), Tectonics, Sedimentary Basins, and Provenance: A Celebration
of William R. Dickinson’s Career. Geological Society of America, Special Paper 540,
https://doi.org/10.1130/2018.2540(09).
PT
Garzanti, E., Ghassemi M.R., Limonta, M., Resentini, A., 2019. Provenance of Karakum desert sand
(Turkmenistan): Lithic-rich orogenic signature of central Asian dune fields. Rivista Italiana di
RI
Paleontologia e Stratigrafia, in press.
SC
Gazzi, P., 1966. Le arenarie del flysch sopracretaceo dell’ Appennino modenese: Correlazioni con il flysch di
Monghidoro. Mineralogica Petrographica Acta, 12, 69-97.
NU
Gazzi, P., Zuffa, G.G., 1970. Le arenarie paleogeniche dell’Appennino emiliano. Mineralogica et Petrographica
Acta, 16, 97-137.
MA
Gazzi, P., Zuffa, G.G., Gandolfi, G., Paganelli, L., 1973. Provenance and dispersal of the sands of the Adriatic
beaches between the Isonzo and Foglia mouths: Regional setting. Società Geologica Italiana, Memorie, 12,
1-37.
ED
Gergen, L.D., Ingersoll, R.V., 1986. Petrology and provenance of Deep Sea Drilling Project sand and sandstone
from the north Pacific Ocean and the Bering Sea. Sedimentary Geology, 51, 26-56.
T
Gilbert, C.M., 1954. Sedimentary rocks. In: Williams, H., Turner, F.J., Gilbert, C.M., Petrography. San
EP
Goldich, S.S., 1938. A study in rock weathering. The Journal of Geology, 46, 17-58.
C
Graham, S.A., Ingersoll, R.V., Dickinson, W.R., 1976. Common provenance for lithic grains in Carboniferous
AC
sandstones from Ouachita Mountains and Black Warrior Basin. Journal of Sedimentary Petrology, 46, 620-
632.
Hamilton, W.B., 2011. Plate tectonics began in Neoproterozoic time, and plumes from deep mantle have never
operated. Lithos, 123, 1-20.
Heberer, B., Röser, G., Behrmann, J.H., Rahn, M., Kopf, A., 2010. Holocene sediments from the Southern Chile
Trench: a record of active margin magmatism, tectonics and palaeoseismicity. Journal of the Geological
Society, London, 167, 539-553.
Helmbold, R., van Houten, F.B., 1958. Contribution to the petrography of the Tanner Graywacke. Geological
Society of America Bulletin, 69, 301-342.
ACCEPTED MANUSCRIPT
41
Henares, S., Bloemsma, M.R., Donselaar, M.E., Mijnlieff, H.F., Redjosentono, A.E., Veldkamp, H.G., Weltje,
G.J., 2014. The role of detrital anhydrite in diagenesis of aeolian sandstones (Upper Rotliegend, The
Netherlands): Implications for reservoir-quality prediction. Sedimentary Geology, 314, 60-74.
Huckenholz, H.G., 1963. Mineral composition and texture in graywackes from the Harz Mountains (Germany)
and in arkoses from the Auvergne (France). Journal of Sedimentary Petrology, 33, 914-918.
Ingersoll, R.V., 1983. Petrofacies and provenance of Late Mesozoic forearc basin, northern and central
California. American Association of Petroleum Geologists Bulletin, 67, 1125-1142.
PT
Ingersoll, R.V., 1990. Actualistic sandstone petrofacies: discriminating modern and ancient source rocks.
Geology, 18, 733-736.
RI
Ingersoll, R.V., 2012. Composition of modern sand and Cretaceous sandstone derived from the Sierra Nevada,
California, USA, with implications for Cenozoic and Mesozoic uplift and dissection. Sedimentary Geology,
SC
280, 195-207.
Ingersoll, R.V., Suczek, C.A., 1979. Petrology and provenance of Neogene sand from Nicobar and Bengal Fans,
NU
DSDP Sites 211 and 218. Journal of Sedimentary Petrology, 49, 1217-1228.
Ingersoll, R.V., Eastmond, D.J., 2007. Composition of modern sand from the Sierra Nevada, California, U.S.A.:
MA
Ingersoll, R.V., Bullard, T.F., Ford, R.L., Grimm, J.P., Pickle, J.D., Sares, S.W., 1984. The effect of grain size
ED
on detrital modes: A test of the Gazzi-Dickinson point-counting method. Journal of Sedimentary Petrology
54, 103-116.
T
Ingersoll, R.V., Cavazza, W., Graham, S.A. and Indiana University graduate field seminar participants, 1987.
EP
Ingersoll, R.V., Kretchmer, A.G., Valles, P.K., 1993. The effect of sampling scale on actualistic sandstone
AC
Johnsson, M.J., 1993, The system controlling the composition of clastic sediments. In: Johnsson, M.J., Basu, A.,
(Eds.), Processes Controlling the Composition of Clastic Sediments. Geological Society of America,
Special Paper 284, 1-19.
Kay, M., 1951. North American geosynclines. Geological Society of America, Memoir 48, 143 p.
Klein, G.deV., 1963. Analysis and review of sandstone classifications in the North American geological
literature, 1940-1960. Geological Society of America Bulletin, 74, 555-576.
Korenaga, J., 2013. Initiation and evolution of plate tectonics on Earth: theories and observations. Annual
Review of Earth and Planetary Sciences, 41, 117-151.
ACCEPTED MANUSCRIPT
42
Krynine, P.D., 1948. The megascopic study and field classification of sedimentary rocks. The Journal of
Geology, 56, 130-165.
Krynine, P.D., 1956. Alice in Graywackeland. Journal of Sedimentary Petrology, 26, 188-189.
Lasius, G., 1789. Beobachtungen über die Harzgebirge, nebst einem Profilrisse, als ein Beytrag zur
mineralogischen Naturkunde. Hannover, Helwingischen Hofbuchhandlung, 132-152.
Limonta, M., Garzanti, E., Resentini, A., Andò, S., Boni, M., Bechstädt, T., 2015. Multicyclic sediment transfer
along and across convergent plate boundaries (Barbados, Lesser Antilles). Basin Research, 27, 696-713.
PT
Mack, G.H., 1984. Exceptions to the relationship between plate tectonics and sandstone composition. Journal of
Sedimentary Petrology, 54, 212-220.
RI
Marsaglia, K.M., 1993. Basaltic island sand provenance. In: Johnsson, M.J., Basu, A., (Eds.), Processes
Controlling the Composition of Clastic Sediments. Geological Society of America. Special Paper 284, 41-
SC
65.
Marsaglia, K.M., 2004. Sandstone detrital modes support Magdalena fan displacement from mouth of Gulf of
NU
California. Geology, 32, 45-48.
Marsaglia, K.M., Ingersoll, R.V., 1992. Compositional trends in arc-related, deep-marine sand and sandstone: a
MA
Marsaglia, K.M., Ingersoll, R.V., Packer, B.M., 1992. Tectonic evolution of the Japanese Islands as reflected in
modal compositions of Cenozoic forearc and backarc sand and sandstone. Tectonics, 11, 1028-1044.
ED
Marsaglia, K.M., Torrez, X.V., Padilla, I., Rimkus, K.C., 1995. Provenance of Pleistocene and Pliocene sand
and sandstone, ODP Leg 141, Chile Margin. Proceedings of the Ocean Drilling Program, Scientific
T
Mawe, J., 1818. A new descriptive catalogue of minerals, consisting of more varieties than heretofore published
and intended for the use of students who may arrange the specimens they collect. London (UK), Longman,
C
McBride, E.F., 1963. A classification of common sandstones. Journal of Sedimentary Petrology, 33, 664-669.
Molinaroli, E., Blom, M., Basu, A., 1991. Methods of provenance determination tested with discriminant
function analysis. Journal of Sedimentary Petrology, 61, 900-908.
Muhs, D.R., 2004. Mineralogical maturity in dunefields of North America, Africa and Australia .
Geomorphology, 59, 247-269.
Okada, H., 1966. Non-greywacke "turbidite" sandstones in the Welsh geosyncline. Sedimentology, 7, 211-232.
Okada, H., 1971. Classification of sandstone: analysis and proposal. The Journal of Geology, 79, 509-525.
ACCEPTED MANUSCRIPT
43
Oriel, S.S., 1949. Definitions of arkose. American Journal of Science, 247, 466-476.
Packer, B.M., Ingersoll, R.V., 1986. Provenance and petrology of Deep Sea Drilling Project sands and
sandstones from the Japan and Mariana forearc and backarc regions. Sedimentary Geology, 51, 5-28.
Packham, G.H., 1954. Sedimentary structures as an important factor in the classification of sandstones.
American Journal of Science, 252, 466-476.
Pettijohn, F.J., 1948. A preface to the classification of the sedimentary rocks. The Journal of Geology, 56, 112-
117.
PT
Pettijohn, F.J., 1949. Sedimentary Rocks. Harper and Brothers, New York, 526 p.
Pettijohn, F.J., 1954. Classification of sandstones. The Journal of Geology, 62, 360-365.
RI
Picard, M.D., McBride, E.F., 2007. Comparison of river and beach sand composition with source rocks,
SC
Dolomite Alps drainage basins, northeastern Italy, In: Arribas, J., Johnsson, M.J., Critelli, S. (Eds.),
Sedimentary Provenance and Petrogenesis: Perspectives from Petrography and Geochemistry. Geological
Society of America, Special Paper 420, 1-12.
NU
Pickering, K.T., Pouderoux, H., Carter, A., Andò, S., Garzanti, E., Limonta, M., Vezzoli, G., Milliken, K.L.,
Chemale, F., Mukoyoshi, H., Kutterolf, S., 2018. Sediment provenance and depositional history of the
MA
Nicobar Fan (Bengal Depositional System) from IODP Expedition 362: detrital zircon geochronology,
apatite thermochronometry, sand petrography and heavy-mineral results. American Geophysical Union
2018, Fall Meeting, T23C-0388.
ED
Potter, P.E., 1978. Petrology and chemistry of modern big river sands. The Journal of Geology, 86, 423-449.
Prasad, S., Hesse, R., 1982. Provenance of detrital sediments from the Middle America Trench transect off
T
Guatemala, Deep Sea Drilling Project Leg 67. Initial reports of the Deep Sea Drilling Project, 67, 507-514.
EP
Rittner, M., Vermeesch, P., Carter, A., Bird, A., Stevens, T., Garzanti, E., Andò, S., Vezzoli, G., Dutt, R., Xu,
Z., Lu, H., 2016. The provenance of Taklamakan desert sand. Earth and Planetary Science Letters, 437,
C
127-137.
AC
Rodgers, J., 1950. The nomenclature and classification of sedimentary rocks. American Journal of Science,
248, 297-311.
Rogers, N.W., James, D., Kelley, S.P., De Mulder, M., 1998. The generation of potassic lavas from the eastern
Virunga Province, Rwanda. Journal of Petrology, 39, 1223-1247.
Schnoor, J.L., Stumm, W., 1986. The role of chemical weathering in the neutralization of acidic deposition.
Swiss Journal of Hydrology, 48, 171-195.
Scholle, P.A., 1979. A Color Illustrated Guide to Constituents, Textures, Cements and Porosities of Sandstones
and Associated Rocks. American Association of Petroleum Geologists, Memoir 28, 201 p.
ACCEPTED MANUSCRIPT
44
Sciunnach, D., Garzanti, E., 2012. Subsidence history of the Tethys Himalaya. Earth-Science Reviews, 25, 179-
198.
Sedgwick, A., Murchison, R.I., 1839. Classification of the older stratified rocks of Devonshire and Cornwall.
The London and Edinburgh Philosophical Magazine and Journal of Science, 14 (89), 241-260.
Singh, S.K., France-Lanord, C., 2002. Tracing the distribution of erosion in the Brahmaputra watershed from
isotopic compositions of stream sediments. Earth and Planetary Science Letters, 202, 645-662.
Sorby, H.C., 1880. On the structure and origin of non-calcareous stratified rocks. Proceedings of the Geological
PT
Society London, 36, 46-92.
Spencer, D.W., 1963. The interpretation of grain size distribution of curves of clastic sediments. Journal of
RI
Sedimentary Petrology, 33, 180-190.
Stewart, R.J., 1977. Neogene turbidite sedimentation in Komandorskiy basin, western Bering Sea. American
SC
Association of Petroleum Geologists Bulletin, 61, 192-206.
Stewart, R.J., 1978. Neogene volcaniclastic sediments from Atka basin, Aleutian Ridge. American Association
NU
of Petroleum Geologists Bulletin, 62, 87-97.Stumm, W., Morgan, J.J., 1996. Aquatic Chemistry, John
Wiley & Sons. Inc., New York.
MA
Suczek, C.A., Ingersoll, R.V., 1985. Petrology and provenance of Cenozoic sand from the Indus Cone and the
Arabian Basin, DSDP Sites 221,222, and 224. Journal of Sedimentary Petrology, 55, 340-346.
Suttner, L.J., Basu, A., 1985. The effect of grain size on detrital modes: a test of the Gazzi-Dickinson point-
ED
Suttner, L.J., Basu, A., Mack, G.H., 1981. Climate and the origin of quartz arenites. Journal of Sedimentary
T
Ukstins, I.A., Renne, P.R., Wolfenden, E., Baker, J., Ayalew, D., Menzies, M., 2002. Matching conjugate
40
volcanic rifted margins: Ar/39 Ar chrono-stratigraphy of pre-and syn-rift bimodal flood volcanism in
C
Ethiopia and Yemen. Earth and Planetary Science Letters, 198, 289-306.
AC
Thornburg, T.M., Kulm. L.D., 1987. Sedimentation in the Chile trench: petrofacies and provenance. Journal of
Sedimentary Petrology, 57, 55-74.
van Andel. T.H., 1958. Origin and classification of Cretaceous, Paleocene and Eocene sandstones of western
Venezuela. American Association of Petroleum Geologists Bulletin, 42, 734-763.
Velbel, M.A., 1985. Mineralogically mature sandstones in accretionary prisms. Journal of Sedimentary
Petrology, 55, 685-690.
Velbel, M.A., 1993. Constancy of silicate-mineral weathering-rate ratios between natural and experimental
weathering: Implications for hydrologic control of differences in absolute rates. Chemical Geology, 105,
89-99.
ACCEPTED MANUSCRIPT
45
Vezzoli, G., Garzanti, E., Vincent, S.J., Andò, S., Carter, A., Resentini, A., 2014. Tracking sediment provenance
and erosional evolution of the western Greater Caucasus. Earth Surface Processes and Landforms, 39,
1101-1114.
Vezzoli, G., Garzanti, E., Limonta, M., Andó, S., Yang, S., 2016. Erosion patterns in the Changjiang (Yangtze
River) catchment revealed by bulk-sample versus single-mineral provenance budgets. Geomorphology,
261, 177-192.
Yerino, L.N., Maynard, J.B., 1984. Petrography of modern marine sands from the Peru‐Chile Trench and
adjacent areas. Sedimentology, 31, 83-89.
PT
Weltje, G.J., 2004. A quantitative approach to capturing the compositional variability of modern sands.
RI
Sedimentary Geology, 171, 59-77.
Weltje, G.J., 2006. Ternary sandstone composition and provenance: An evaluation of the ‘Dickinson model’. In:
SC
Buccianti, A., Mateu-Figueras, G., Pawlowsky-Glahn, V. (Eds.), Compositional data analysis: From theory
to practice. Geological Society of London, Special Publication 264, 611-627.
NU
Weltje, G.J., 2012. Quantitative models of sediment generation and provenance: state of the art and future
developments. Sedimentary Geology, 280, 4-20.
MA
Weltje, G.J., von Eynatten, H., 2004. Quantitative provenance analysis of sediments: Review and outlook.
Sedimentary Geology, 171, 1-11.
Werner, A.G., 1787. Kurze Klassifikation und Beschreibung der verschiedenen Gebürgsarten. Dresden,
ED
Waltherischen Hofbuchhandlung, 28 p.
Whetten, J.T., 1969. Sediments from the lower Columbia River and origin of graywacke. Science, 152, 1057-
T
1058.
EP
Whetten, J.T., Hawkins, J.W., 1970. Diagenetic origin of graywacke matrix minerals. Sedimentology, 15, 347-
361.
C
Whetten, J.T., Kelley, J.C., Hanson, L.G., 1969. Characteristics of Columbia River sediment and sediment
AC
White, N., Pringle, M., Garzanti, E., Bickle, M., Najman, Y., Chapman, H., Friend, P., 2002. Constraints on the
exhumation and erosion of the High Himalayan slab, NW India, from foreland basin deposits. Earth
Planetary Science Letters, 195, 29-44.
Wolf, K.H., 1965. Gradational sedimentary products of calcareous algae. Sedimentology, 5, 1-37.
Wolf, K.H., 1971. Textural and compositional transitional stages between various lithic grain types (with a
comment on “Interpreting detrital modes of graywacke and arkose”). Journal of Sedimentary Petrology, 41,
328-332.
ACCEPTED MANUSCRIPT
46
Zuffa, G.G., 1980. Hybrid arenites: Their composition and classification. Journal of Sedimentary Petrology, 50,
21-29.
Zuffa, G.G., 1985. Optical analyses of arenites: influence of methodology on compositional results. In: Zuffa,
G.G. (Ed.), Provenance of arenites. Reidel, Dordrecht, NATO ASI Series 148, 165-189.
Zuffa, G.G., 1987. Unravelling hinterland and offshore palaeogeography from deep-water arenites. In: Leggett,
J.K., Zuffa, G.G. (Eds.), Marine clastic sedimentology. Concepts and case studies. Graham and Trotman,
London, 39-61.
PT
Zuffa, G.G., Normark, W.R., Serra, F., Brunner, C.A., 2000. Turbidite megabeds in an oceanic rift valley
recording jökulhlaups of late Pleistocene glacial lakes of the western United States. The Journal of
RI
Geology, 108, 253-274.
SC
NU
MA
T ED
C EP
AC
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8