Full text of DOI:10.1080/030094801750203170

QuantiŽ cation of SEM microtextures useful in sedimentary
environmental discrimination
WILLIAM C. MAHANEY, ANDREW STEWART AND VOLLI KALM
Mahaney, W. C., Stewart, A. & Kalm, V. 2001 (June): QuantiŽ cation of SEM microtextures useful in sedi-
mentary environmental discrimination. Boreas, Vol. 30, pp. 165–171. Oslo. ISSN 0300-9483.
Scanning electron microscopic imagery is often used to identify and discriminate among environments of sedi-
mentation with the main aim of identifying individual microfeatures, or suites of microtextures, that are con-
sidered indicative of a particular depositional environment or geologic process. Because few microtextures are
considered to represent a single geologic process it is necessary to analyze a large number of quartz sands and
other mineralic grains with the objective of determining the frequency of occurrence of a range of microtex-
tures within a distinct sample suite. Using percent frequency of occurrence of different microtextures from
suites of  uvial, glacio uvial and glacial sands from sites in Estonia and Latvia, we invoked statistical com-
parison of different sample suites using Euclidean distances. These provide a quantitative means of measuring
the differences among different sediments and processes that formed them and also a quantiŽ cation tool useful
in assessing microtextures as a recorder of sedimentary environments.
William C. Mahaney (e-mail: bmahaney@yorku.ca) and Andrew Stewart, Geomorphology and Pedology
Laboratory, York University, 4700 Keele St., North York, Ontario, Canada, M3J 1P3; Volli Kalm (e-mail:
vkalm@math.ut.ee), Tartu University, Institute of Geology, 46 Vanemuise Str., 51014 Tartu, Estonia; received
13th July 2000, accepted 26th October 2000
Recent research on discrimination of quartz sands from sedimentary environments from the range of imprinted
 uvial, glacio uvial and glacial sediments in Estonia microtextures. Some researchers have pointed to evi-
and Latvia (Mahaney & Kalm 2000) has relied on bar dence that some microtextures could be produced by all
graph signatures for each sample suite to demonstrate geologic agents (see, for example, Brown 1973), which
differences between individual depositional environ- eventually resulted in the principle of equiŽ nality, i.e.,
ments (Fig. 1). The bar graphs indicate total percentage different geologic processes could produce similar
of occurrence of each microtexture in each of three microfeatures on sand size clasts. Still other workers
sampled environments: (1) indurated and non-cemented criticized operator variance as a factor important in
¨
sandstones of Middle Devonian (Eifel) Arukula Stage producing different interpretations from similar groups
sandstone deposited in a  uvial environment; (2) quartz of samples; others believe that bedrock source plays a
sands from basal till deposited under thick Weichselian more important role than process and depositional
ice; and (3) glacio uvial sands deposited by meltwater environment in producing the range of microtextures
under stagnant ice or around an ice margin (Mahaney & seen on groups of grains (Mazzulo & Ritter 1991).
Kalm 2000). Using replicate samples from each site, the
Because many SEM studies have relied on sample
objective of Mahaney & Kalm (2000) was to determine suites collected close to glacial environments, it is
the range and frequency of microtextures present in the entirely possible that many grains in non-glacial
three groups of samples and to determine differences sediments might have been transported through the
among them. The location of sites, methods employed glacial mill. With this in mind, it is logical to assume
and stratigraphy, particle size data and SEM imagery that grains might often be reworked from glacial
are interpreted along with the fractography in Mahaney environments and mixed with aeolian,  uvial and
& Kalm (2000). In this article, we attempt to deŽ ne, coastal sediments (see e.g. Setlow & Karpovich
more rigorously, the statistical differences among these 1972). In studies of aeolian grains (Mahaney & Andres
three sample groups in terms of the suite of micro- 1996), grains with predominantly low relief, minimum
textures present in each using Euclidean differences.
parallel and conchoidal fractures and high percentages
of abrasion microfeatures characteristic of wind-blown
transport were contrasted with grains derived from the
Ordovician diamictite of northern Algeria. The recov-
ered quartz sands of glacial origin carried much greater
percentages of deeply imbedded fracture surfaces,
higher relief, grooves (striations) and step features.
In Estonia and Latvia, the study of indurated and non-
cemented sandstones of Middle Devonian (Eifel)
Background
As pointed out by Mahaney & Kalm (2000), since the
early years of SEM study of sand grain microtextures
(Krinsley & Takahashi 1962; Krinsley & Doornkamp
1973; Whalley & Krinsley 1974; Bull 1981; Perttunen
& Hirvas 1982; Mahaney & Krinsley 2001), the main
objective has been, and still is, the identiŽ cation of
¨
Arukula Stage sandstone deposited in a  uvial environ-
ment were compared with quartz sands from basal till

166
William C. Mahaney et al.
BOREAS 30 (2001)
Fig. 1. Percent occurrence of microtextures in three environments: A, Devonian sands; B, Weichselian tills; and C, Weichselian glacio u-
vial sands (from Mahaney & Kalm 2000).

BOREAS 30 (2001)
QuantiŽ cation of SEM microstructures
167
Fig. 2. Geological map of Estonia and part of northern Latvia showing the position of the Rann section west of Tallinn along the northern
coast and the transect from which samples of Weichselian Till and Devonian Sands were collected (see Mahaney & Kalm (2000) for loca-
tions of till/sandstone sections). The explanations of lithological units are: PR = Proterozoic granitic and metamorphic rocks of the Fen-
noscandian Shield; C = Precambrian and Cambrian claystones, sandstones and siltstones; O = Ordovician limestones and dolomites;
S = Silurian dolomites and limestones; D = Devonian sandstones; UD = Upper Devonian dolomites. Arrows indicate the  ow vectors of
Weichselian ice.
deposited under thick Weichselian ice (Mahaney & the Fenno-Scandian Shield 200 km to the north. Out-
Kalm 2000) and glacio uvial sands deposited by crops of Cambrian sandstone along the northern
meltwater under the stagnant ice or around the ice Estonian coast and along the seabed of the Gulf of
margin. Using replicate samples, the objective of Finland contain only very Ž ne sand and coarse silt
Mahaney & Kalm (2000) was to determine the (Kurvits et al. 2000) and thus could not have con-
difference, if any, of the microtextures present on sands tributed sands above 100 mm to Quaternary glaciers.
in three groups of samples. The methods employed and
the stratigraphy, particle size data and SEM imagery are
interpreted along with the fractography in Mahaney &
Kalm (2000).
Data summary
Excluding the Rann section (Mahaney & Kalm A summary of the imagery presented in Mahaney &
2000), located on Cambrian sandstone along the Gulf Kalm (2000) is shown in Figs 3 and 4. Grains subjected
of Finland, the sample transect stretches from the shores
to percussion-strength contact in a  uvial system are
of Lake Peipsi in the northeast of Estonia to the shown in Fig. 3. The general frame (A) shows 21
southwest, where it ends near Ruhja in Latvia. It
traverses a leg c. 80 km long that includes till outcrops
rounded to subrounded grains; the remainder classify as
subangular. The most rounded grains in the Devonian
overlying Devonian sandstone; sampled sites are shown sandstone are presumably those clasts that have under-
in Fig. 2. Many of the quartz grains larger than 100 mm, gone the greatest distance of transport and contact with
in till, came either from the Devonian outcrops or from
each other in turbulent water. These grains have also

168
William C. Mahaney et al.
BOREAS 30 (2001)
Fig. 3. A. General frame of sediments in sample TAR3-3 showing a range of angular to rounded grains from the Devonian sandstone. B.
Fluvial sand from site TAR2-4 with bulbous protrusions indicating a wind-blown history and a multitude of v-shaped percussion cracks,
dissolution etching and craters of unknown origin. C. Quartz subrounded sand from sample TAR1-3 with rounded edges, v-shaped percus-
sion cracks, large craters (side) and dissolution etching. D. Nearly rounded  uvial grain from sample TAR3-3 with numerous v-shaped
percussion cracks, minor craters and dissolution etching surrounded by subrounded to subangular grains.
endured the greatest degree of reworking and also carry for dissolution, precipitation and coatings compared
the highest counts of v-shaped percussion scars (Figs with grains recovered from tills (Fig. 4A–D).
3B, D). A count of weathered to fresh v-shaped
Quartz grains recovered from till (Fig. 4A) are less
percussion scars yields a value of 45% weathered to rounded (viz. Fig. 3A vs. 4A), give a ratio of 11/21 of
55% fresh, which indicates that many percussion crack subangular to subrounded grains. Fig. 4A shows a
microtextures are altered post-depositionally.
greater number of grains with triangular-faceted shapes
In some cases the presence of bulbous edges (Fig. 3B) carrying a range of deeply imbedded fracture patterns in
indicates the addition of aeolian grains, and in still other addition to straight and curved grooves or striations
cases weathering features (Fig. 3C) highlight diagenetic (Fig. 4C). Relief is moderate to extreme (Fig. 4C, D) on
dissolution, partially interconnected with v-shaped 50% of the grains sampled in contrast to grains from the
fractures. These may be initial planes of weakness  uvial sample suite which exhibit generally subdued
that are preferentially attacked by ground or meteoric relief.
waters invading the sandstone (see Kurvits et al. 2000).
The degree of relief, seen to be extreme on glacial
Examination of microtextures on sands recovered grains (Perttunen & Hirvas 1982; Helland & Diffendal
from Devonian outcrops (as shown in Fig. 3) depicts the 1993; Mahaney 1990, 1995; Mahaney et al. 1988, 1996;
 uvial sands as grains of low relief, with moderate Mahaney & Kalm 1995, 2000; Mahaney 2001), is
display of fractures, lacking grooves, step or chatter- without doubt due to grain-to-grain contact in the basal
mark features and carrying considerably more evidence ice under extreme shear stress. In contrast, some grains

BOREAS 30 (2001)
QuantiŽ cation of SEM microstructures
169
Fig. 4. A. General frame of angular to subrounded grains with many  akes of mica in sample TAR4-5 showing mainly angular to suban-
gular grains. B. Quartz sand with a pre-crushed surface showing multiple v-shaped percussion scars, small crater on left  ank overprinted
with a large crater and subparallel fractures in till RUH1-2. C. Angular quartz with sharp edges, high relief, striations (lower area),
abraded fractures in till Vilandi1-3. D. Angular quartz with maximum relief in till Elva1-3 showing sharp edges, crater, subparallel frac-
tures, single step features and silts packed and indurated on step above center. Local till names from Mahaney & Kalm (2000).
complete a sojourn in the ice without contacting other
grains and these may retain their original depositional
environmental signatures. Others (Fig. 4B) may be
Analysis of quantiŽ cation data on
microtextures
partially reformed and still retain a preglacial record, Bar graph summaries of microtexture observations on
which often may include some preweathering (Vortisch
sediment grains from three environments of deposition
– Devonian  uvial sands, Weichselian glacio uvial
et al. 1987).
Because quartz grains in the Devonian sandstone sands and Weichselian tills – previously published
could have an origin in the Precambrian rocks of the
(Mahaney & Kalm 2000) are reproduced here (Fig. 1)
Fenno-Scandian Shield, from the Cambrian sandstones for further analysis. The purpose of this analysis is to
of northern Estonia and the Gulf of Finland, and/or from order the depositional environments based on summary
Devonian outcrops in southern Estonia and northern
microtexture observations to determine whether order-
Latvia, it is impossible to determine precisely the ing re ects similarities or relationships among distinct
distance of transport by Weichselian ice. However,
coarse quartz grains with diameters greater than 100 mm
geological processes –  uvial, glacio uvial and glacial.
Vertical axes (Fig. 1) in the bar graphs represent
must come from the Devonian outcrops, or from the percentage of observations – and the percentage of
Baltic Shield, as the Cambrian sands are limited to very grains observed in that environment which exhibit a
Ž ne sand and coarse silt (Kurvits et al. 2000).
particular microtexture. All microtextures are repre-

170
William C. Mahaney et al.
BOREAS 30 (2001)
Table 1. Euclidean distances between different groups of sedi-
possible to demonstrate this one way or the other. Two-
way plots suggest that some variables are probably
related, among them: edge-rounding and sharp angular
features (negatively); preweathered surfaces and
dissolution etching (positively); and overprinted grains,
conchoidal fractures and adhering particles (all posi-
tively with each other). Quantitative analysis of
observations on individual sediment grains from all
three environments would be required to show these
relationships, which is beyond the scope of this paper.
Distance values conŽ rm the impression that  uvial
sands are distinct from tills. They provide additional
ments.
Weichselian
Weichselian glacio uvial
Devonian
 uvial sands tills
sands
Devonian  uvial sands
Weichselian tills
Weichselian
0
145
77
0
111
0
glacio uvial sands
sented in each graph, arranged in the same order along information, allowing the three environments to be
their horizontal axes. ordered in a way that makes intuitive sense: Weichse-
The bar graphs have different shapes, suggesting that lian tills and Devonian  uvial sands are most distant
there are distinct signatures for each environment. More (145), occurring at opposite ends of a spectrum;
precisely, the two sands ( uvial and glacio uvial) seem glacio uvial sands are intermediate, being closely
to resemble each other more than either resembles till. related to  uvial sands (77) and, more distantly, to tills
This is apparent in the range of microtextures observed; (111). This latter relationship suggests that glacial
–  uvial sands are represented by fewer (15) and tills by grains subjected to meltwater transport may be quickly
more (23) microtextures. Moreover, the speciŽ c micro- reformed with greater relief and a multitude of v-shaped
textures present and absent on sands are the same ones percussion cracks, the latter considered to be the
in each case, except for ‘sharp angular features’, which deŽ nitive recorder of turbulent water transport.
are present on glacio uvial but absent on  uvial sands.
This Ž rst approximation of grouping – sands distinct
from tills – can be tested and reŽ ned by application of a
distance or similarity coefŽ cient.
Conclusion
The idea that  uvial and glacio uvial sands are more
like each other than they are to tills can be tested by
ordering these environments using a coefŽ cient that
re ects the distance (or similarity) among them based
on percentages of microtextures. Euclidean distances
(shown in Table 1) re ect straight-line distances
between all possible pairs of environments based on
the equation:
While much can be inferred from the percentage data
presented in bar graph form, a quantitative test of the
distance (or similarity) among the percentages of
microtextures underscores the important differences
between quartz particles transported by water and ice.
Moreover, the statistical measure shows the importance
of grain overprinting, which occurs; for instance, when
grains start in the glacial system and become over-
printed with  uvial microtextures during transport in
meltwater. Euclidean distance measures suggest that,
while glacial grains retain damage in icted upon them
during their time in the ice, these grains slowly reform
with meltwater transport and take on the surface
morphometry of  uvial clasts.
s
p
X
2
dij ˆ
…xik ¡ xjk†
kˆ1
where the distance, d, between two cases (environ-
ments), i and j, is measured in terms of a number of
variables (microtextures), p. This coefŽ cient is com-
monly used to measure distance between cases or
objects which have been measured on a number of
metrical variables. The distance coefŽ cient has no
restricted range (between zero and inŽ nity) and
measured distances between pairs of cases are relative,
that is, values in Table 1 are useful only for comparison
to each other.
Usually, two assumptions are made when interpreting
results (Shennan 1988). First, the range in values for all
of the original variables (Fig. 1) should be similar –
here, all variables are percentages and so values range
from 0 to 100 (as shown in Fig. 1), although it is rare to
Ž nd values greater than 60%. Second, variables should
be independent. Here, it is likely that some variables are
dependent on others, but with only three cases it is not
References
Brown, J. E. 1973: Depositional histories of sand grains from sur-
face textures. Nature 242, 396–398.
Bull, P. A. 1981: Environmental reconstruction by scanning elec-
tron microscopy. Progress in Physical Geography 5, 368–397.
Helland, P. E. & Diffendal, R. F., Jr. 1993: Probably glacial cli-
matic conditions in source areas during deposition of parts of
the Ash Hollow Formation, Ogallala Group (Late Tertiary) of
western Nebraska. American Journal of Science 293, 744–757.
Krinsley, D. H. & Takahashi, T. 1962: The surface textures of
sand grains: an application of electron microscopy. Science 135,
923–925.
Krinsley, D. H. & Doornkamp, J. C. 1973: Atlas of Sand Grain
Surface Textures. 91 pp. Cambridge University Press, Cam-
bridge, U.K.
Kurvits, T., Mahaney, W. C. & Kalm, V. 2000: Grain micromor-

BOREAS 30 (2001)
QuantiŽ cation of SEM microstructures
171
phological characteristics in the Rannamoisa Section, Cambrian
sandstones, northern Estonia. Proceedings Estonian Academy of
Sciences, Geology 49, 17–26.
Mahaney, W. C. 1990: Ice on the Equator. 386 pp. Wm. Caxton
Ltd., Ellison Bay, Wisconsin.
and continental ice – some examples from North America and
East Africa. American Journal of Science 288, 810–826.
Mahaney, W. C., Campbell, I. A. & Claridge, G. 1996: Microtex-
tures on quartz grains in tills from Antarctica. Palaeogeography,
Palaeoclimatology and Palaeoecology 121, 89–103.
Mahaney, W. C. 1995: Pleistocene and Holocene glacier thick-
nesses and/or transport histories inferred from microtextures on
quartz particles. Boreas 24, 293–304.
Mazzullo, J. & Ritter, C. 1991: In uence of sediment source on
the shapes and surface textures of glacial quartz sand grains.
Geology 19, 384–388.
Mahaney, W. C. 2001: Atlas of Sand Grain Surface Textures and Perttunen, M. & Hirvas, H. 1982: An attempt to use the roundness
Applications. 256 pp. Oxford University Press, Oxford, U.K. In
press.
Mahaney, W. C. & Andres, W. 1996: Scanning electron micros- Setlow, L. M. & Karpovich, R. P. 1972: ‘Glacial’ microtextures on
copy of quartz sand from the north-central Saharan desert of
of quartz grains for till stratigraphy. Bulletin Geological Survey
of Finland 54, 25–33.
quartz and heavy mineral sand grains from the littoral environ-
ment. Journal of Sedimentary Petrology 42, 864–975.
Shennan, S. 1988: Quantifying Archaeology. 364 pp. Edinburgh
University Press, Edinburgh, Scotland.
Vortisch, W. B., Mahaney, W. C. & Fecher, K. 1987: Lithology
and weathering in a paleosol sequence on Mount Kenya. Geolo-
gica et Paleontologica 15, 67–91.
Whalley, W. B. & Krinsley, D. H. 1974: Scanning electron micro-
scope study of surface textures of quartz grains from glacial
environments. Sedimentology 21, 87–105.
¨
Algeria. Zeitschrift fur Geomorphologie, N.F. Supplement-Bd.
103, 179–192.
Mahaney, W. C. & Kalm, V. 1995: Scanning electron mi-
croscopy of Pleistocene tills in Estonia, Boreas 24, 13–29.
Mahaney, W. C. & Kalm, V. 2000: Comparative SEM study of
oriented till blocks, glacial grains and Devonian sands in Estonia
and Latvia. Boreas 29, 35–51.
Mahaney, W. C., Vortisch, W. & Julig, P. 1988: Relative differ-
ences between glacially crushed quartz transported by mountain

Copyright of Boreas is the property of Wiley-Blackwell and its content may not be copied or
emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.