:4... :1 . ,
.w . ,
. e 3..
, , , .r‘m

1.... .

w” ééymgfi g

m. m. .

 

.

 

 

 

. ,-
llh a! 3“-

$95...

113‘,"
ifiun..hrlh.u.fitha Iron
3&3...» 3.1....»
s at}. at!
2...

i...
I I» z:
1.53).!!!
.: 1.... 3 If...‘
r. .15....iu: 2”. lit...
1 L .152. {is

a 910.13....) at.
1. :9.)
3‘ N: 5:). 15.3...
d. It" 411. ix.
. (it 1113.1 i} I
ll.‘slh!\
5.}.335’ s
it’llu‘JI x.
‘3 . (ii fit...
1 «a:
.1; 32!!

. . .u .1. EL:
gig!!! .51.? .
ii..- 5.3.3:... 1i.
4.39.1.1! Entity: .34!
r 9.581 :31...-

“fit-‘55:...
.Iiiil: .
1‘ 515’. '3‘»: \‘Y.
.1 ‘Lf‘.}(.)1on?9)

1'. a." .
if...)
L. o pl

 

kw ‘ 4%. $3..
.. . . QLWMX.

 

 

 

,2 ,
10m LISP " nnY
Michigan State
University

 

 

This is to certify that the
thesis entitled

CHARACTERISTICS AND ORIGIN OF SILTY KETTLE BOTTOM
DEPOSITS IN A SANDY NORTHERN LOWER MICHIGAN
LANDSCAPE

presented by

TREVOR C. HOBBS

has been accepted towards fulfillment
of the requirements for the

MS. degree in Geography

 

 

 

Major Profesfis Signature

41-1-0?

 

Date

MSU is an Affirmative Action/Equal Opportunity Employer

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:/Proj/Aoc&Pres/ClRC/DateDue,indd

 

1-

. “if“! .‘r.’"'.'.“‘“‘.-" 1‘12”“: (“Jr-"3 " “3"

muilliifi‘iii

=fl"--i_‘m m

 

 

‘ “a: .m—.- a...” m. .—.._._._“ M

CHARACTERISTICS AND ORIGIN OF SILTY KETTLE BOTTOM DEPOSITS IN A
SANDY NORTHERN LOWER MICHIGAN LANDSCAPE

By

Trevor C. Hobbs

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Geography

2009

ABSTRACT

CHARACTERISTICS AND ORIGIN OF SILTY KETTLE BOTTOM DEPOSITS IN A
SANDY NORTHERN LOWER MICHIGAN LANDSCAPE

By
Trevor C. Hobbs

Many dry, topographically closed basins (i.e. kettles) in the sandy interlobate
uplands of Osceola county contain a ~1m thick sequence of silty sediment in the bottom-
center of the depression. These deposits are anomalous, when compared to the
preponderance of sand in this area. The purpose of this research was to determine the
most likely origin(s) of these silty deposits, and assess their paleoenvironmental
importance. Two possible hypotheses explaining the origin of the silt were examined: (1)
it was winnowed out of the surrounding upslope sediments via slopewash, or (2) it is
loess. Sediments from 53 kettle bottoms and adjacent backslopes were sampled to
characterize and compare their particle size distributions. Eight of these kettle bottoms
were also depth-sampled to reconstruct their depositional histories. Silt mineralogy data
from four kettle bottoms and adjacent backslopes were later compared to determine the
' likely source of the silt. Two samples of charcoal fiagments found in buried soils in silty
kettle bottom sediments yielded radiocarbon ages of 6840 2+: 30 and 920 :1: 20 cal. years
BP. The combined evidence suggests that the kettle bottom silt is locally redistributed
loess that was deflated fiom surfaces destabilized by fire during the Holocene and
preferentially deposited in kettle bottoms. These data provide a new and potentially
important paleoenvironmental proxy for recently glaciated landscapes, and highlight the
point that episodic periods of dryness and landscape instability have permeated the post-

glacial period in Michigan.

Dedicated to my loving Grandmother, Lucille, who exemplified perseverance

iii

ACKNOWLEDGEMENTS

I’m fortunate to have learned from those whom I consider to be the best workers
and teachers in the field of geomorphology. First and foremost, this thesis would not
have been possible without the inspiration, guidance, and support of my advisor, Randy
Schaetzl- thank you for everything you’ve given me. Also, my other committee
members, Alan Arbogast and Grahame Larson, deserve special thanks for being role
models and excellent teachers. I’d also like to thank Dave Lusch and Kevin Kincare for
lending their brilliant ideas and time in the field. Finally, I’d like to thank Brandon Curry
and Catherine Yansa for contributing their input.

This thesis would not have been possible without the camaraderie and support of
my fellow graduate students. Thank you Steve for your relationship with Endnote.
Special thanks goes out to Bilal (Yoda) and Meleia for bagels and honest conversation.
Blumer- I could always count on you to have coffee ready every morning. Kristy- thanks
for feeding me various snacks when I was starving. J oel- thanks for showing me new
music. Zukas- thanks for really showing me new music (no offense Joel), and upping the
homebrew quality. Bigsby- thanks for your workspace and setting an example of how to
run far. To all other grad students in the department of Geography- thanks for setting the
bar high.

I am forever indebted to my network of friends in the Lansing community who
have lent unconditional support over the past two years. To Corie Peirce, Chris Donnan,
little H-man, Tomm Becker, and the community of folks at the MSU Student Organic

Farm- thank you for what you do. Your food has nourished my intellect and continually

iv

fuels healthy bodies and conscious minds everywhere. Also, to all those folks in the
Eastside neighborhoods (there are too many of you to list)- thanks for the potlucks,
laughs, and being great neighbors and citizens. To all the amazing housemates I’ve been
fortunate enough to live with and learn from over the past two years- you’ve shaped me;
thank you. Finally, the Waffle house (222 Horton) deserves special thanks for always
keeping an open door.

I am infinitely grateful to the network of musicians and people in the Earthwork
community for balancing out my graduate life. Micah and Andrea- thank you for
maintaining the creative spirit in me, and believing that there would be a day when I
finish my thesis. Chris and Brandon- thank you for opening up to me, both musically and
emotionally. And to all others that I’ve been fortunate enough to play, record, and
associate with over the years- you’ve enriched my graduate experience with your music.

Other people graciously contributed their time and/or resources in the field: to
my best friend J onny B, and students Kelly Anderson, Erin Dixon, Daniel Star, and
Aaron Walker- thank you for your time in the field. To Steph, thanks for letting me drive
your Subaru, and sorry that I destroyed the undercarriage.

Finally, and most importantly, thank you to my parents and family for being my
rock- you’ve always tolerated and believed in my pursuits, and for that I am forever

gratefirl.

TABLE OF CONTENTS

CHAPTER 1 - Introduction ................................................................................................ 1
CHAPTER 2 - Study Area .................................................................................................. 7
2.1 Geographic Setting .................................................................................................... 7
2.2 Study Area Extent ..................................................................................................... 7
2.3 Hydrography ............................................................................................................. 9
2.4 Geomorphology ...................................................................................................... 13
2.5 Soils ......................................................................................................................... 27
2.6 Vegetation ............................................................................................................... 31
2.7 Climate .................................................................................................................... 38
CHAPTER 3 - Methods .................................................................................................... 40
3.1 Preliminary Work .................................................................................................... 40
3.2 Field Methods ......................................................................................................... 41
3.3 Laboratory Methods ................................................................................................ 49
CHAPTER 4 - Results and Discussion ............................................................................. 56
4.1 Extent of Silt in Kettle Bottoms .............................................................................. 57
4.2 Texture of Kettle Bottom Deposits and Surrounding Backslopes .......................... 61
4.3 Comparing Silt Content in Kettle Bottom and Backslope Pairs ............................. 67
4.4 Silt Contents with Depth - Geomorphically Stable Sites ........................................ 82
4.5 Spatial Patterns of Silt Content - Backslopes vs. Kettle Bottoms ........................... 92
4.6 Depth Characteristics of Kettle Bottom Deposits ................................................... 95
4.7 Timing of Silt Deposition ..................................................................................... 119
4.8 Source of Kettle Bottom Silt - Local vs. Extra Regional? .................................... 126
4.9 Origin of Kettle Bottom Deposits - Summary of Evidence .................................. 134
CHAPTER 5 - Paleoenvironmental Significance ........................................................... 136
5.1 Soil Charcoal as an Indicator of Fire .................................................................... 136
5.2 Fire on Sandy Substrates ....................................................................................... 138
5.3 A Warmer and Dryer Climate During the Mid-Holocene .................................... 139
5.4 Evidence for Climate Fluctuation from Nearby Studies ....................................... 140
5.5 The Possible Role of Climate and Fire in Destabilizing Sandy Surfaces ............. 145
5.6 Paleoenvironmental Significance - Summary ....................................................... 146
CHAPTER 6 - Conclusions ............................................................................................ 148
6.1 Geomorphology of the Evart Upland .................................................................... 148

vi

6.2 Characteristics and Origin of Silty Kettle Bottom Deposits ................................. 148

6.3 Silt Deposition Contemporaneous with Holocene Fires ....................................... 151
6.4 Kettle Bottom Silt is Local ................................................................................... 151
6.5 Paleoenvironmental Significance .......................................................................... 152
APPENDIX ..................................................................................................................... 153
BIBLIOGRAPHY ........................................................................................................... 1 58

vii

LIST OF TABLES

Table 2.1- Morphometric characteristics of kettles. ......................................................... 17

Table 2.2— Characterization data for the five most common soil series mapped on the
Evart Upland , listed in order of abundance from most (top) to least (bottom) ........ 28

Table 2.3- Average monthly climate data for 30-year period (1971-2000) ...................... 39
Table 4.1- Average textural components of backslope and kettle bottom deposit samples.
................................................................................................................................... 62
Table 4.2- Downslope, point-to-point comparisons of mode sand grain size. ................. 77
Table 4.3- Summary of evidence regarding the origin of kettle bottom deposits ........... 135
Table A.1- Textural data for backslope samples. Numbers expressed as percent of
sample. MWPS = mean weighted particle size ...................................................... 154
Table A.2- Textural data for kettle bottom deposit samples. Numbers expressed as
percent of sample. MWPS = mean weighted particle size ..................................... 156

viii

LIST OF FIGURES

Figure 1.1- The Evart Upland is a complex assemblage of outwash with many kettles,
situated between the Lake Michigan and Saginaw lobes of the former Laurentide ice
sheet. The Evart Upland is in central Osceola County, MI (highlighted red).

Images in this thesis are presented in color. ............................................................... 2

Figure 2.1- Color, hill-shaded, digital elevation model (DEM) of Osceola County, MI.
The extent of the Evart Upland is shaded gray. The dashed line represents the Pere-
Marquette State Forest boundary. The black lines are major highways. ................... 8

Figure 2.2- Combined topographic map (3-meter contour interval) and color-shaded
DEM of the western portion of the Evart Upland. The black line separates dry
kettles on the upland from bogs, wetlands, and lakes in the lowlands, and is the
approximate boundary of the study area. .................................................................. 10

Figure 2.3- Combined topographic map (3-meter contour interval) and color-shaded
DEM of the northeastern portion of the Evart Upland. Lower relief, wet, hurnmocky
terrain characterizes the surrounding lowlands to the northeast. .............................. 11

Figure 2.4- Grayscale DEM of the Evart Upland and surrounding areas, showing
hydrologic features. Blue lines represent perennial streams and rivers (data source:
Osceola County Hydrography shapefiles from the Michigan Geographic Data
Library (Michigan Department of Information Technology, 2007)). The Muskegon
River, located south of the upland, flows to the southwest. The Hersey River,
flowing south, joins the Muskegon River at the bottom of the map ......................... 12

Figure 2.5- A) Grayscale DEM of the Evart Upland showing the location of transect line
A-B (red) that was used to construct the topographic profile shown in part B. The
NNE trending ridge extending from the eastern portion of the study area is not
considered part of the Evart Upland. B) Topographic profile of transect line A-B.
Kettles on the upland have much greater local relief than kettles in the surrounding
lowlands near the Muskegon River ........................................................................... 16

Figure 2.6- Grayscale DEM of the study area with major kettles highlighted in red. A
total of 535 kettles were mapped within the boundary of the Evart Upland. ........... 18

Figure 2.7- Grayscale DEM showing two areas on the Evart Upland with distinct kettle
morphology (below); one from the western portion (upper left), the other from the
southeastern portion (upper right). The two topographic maps above (3 -meter
contour interval) are the same scale. ......................................................................... 19

Figure 2.8- USGS topographic map (3-meter contour interval) showing examples of
ridges (ridge tops shaded tan) next to kettles (shaded red) in the study area. The

ix

ridges in the left map are relatively flat topped and have steep sides. The ridge in
the right map is massive with deeply gullied sides. .................................................. 21

Figure 2.9- Combined USGS topographic map (3 -meter contour interval) and color-
shaded DEM showing low areas (yellow shading) between ridges (brown shading).
In these low areas kettles are grouped. Major ridge crests are traced with black
lines. Kettles are shaded red ..................................................................................... 22

Figure 2.10— Combined hillshade and color-shaded DEM presented in a three-
dimensional view looking northwest, at the Evart Upland. The boundary of the
upland is outlined in black. The red lines trace weakly integrated stream valleys that
are headed by kettle chains upslope. ......................................................................... 24

Figure 2.11- A) Topographic map (3-meter contour interval) of a portion of the Evart
Upland. The red polygons outline kettles. The blue line outlines the drainage basin
of kettles that subtly transition to Twin Creek, south of the map. Note that the long
axis orientation of the kettles that occupy the central portion of the valley is more or
less parallel to the valley. Kettles in the northern most portion of the drainage basin
are not well integrated. B) Inset map of a grayscale DEM showing the location of
the drainage basin (shaded blue) on topographic map in (A). The Evart Upland is
outlined in black. Blue lines on the inset map are perennial streams. ..................... 25

Figure 212- Color shaded DEM showing examples of a composite kettle (A) and lake
(B) that have contrasting long axis trends as described in the text. Kettles are shaded
red. Lakes are not shaded. ........................................................................................ 27

Figure 213- Map of the five most common soil series on the Evart Upland (SSURGO
digital soils data (Soil Survey Staff, 2007)). All other mapped series are left blank
(gray areas on the upland). Grayling and Rubicon soils are mapped in narrow
valleys, deep, dry kettles, and low areas between ridges. McBride soils increases in

abundance on the east side of the upland, while Rubicon soils are generally absent
there ........................................................................................................................... 30

Figure 2.14 Combined hillshade and presettlement vegetation map showing the Evart
Upland and surrounding areas. The study area boundary is outlined black.
Presettlement vegetation data from (Comer et a1., 1998). ........................................ 33

Figure 2.15- The Evart Upland is primarily used for timber production. Clear cutting is
common. Photo by R. Schaetzl. ................................................................................ 34

Figure 216- Some kettles lack canopy cover. In the photograph above, the surrounding
forested summits open up to sweet fern and grasses (visible in the foreground) in the
kettle bottom. Photo by R. Schaetzl. ......................................................................... 34

Figure 2.17- Topographic map (3-meter contour interval) of the northwestern portion of
the study area. Non-forested kettles are shaded white on this topographic map.

Black arrows point to examples of kettles that are non-forested. Other larger areas
with white shading were clear-cut at the time the topographic map was made (1983).
................................................................................................................................... 35

Figure 2.18- A) Bracken fern is common on steep slopes (above photo looking up a kettle
backslope with a 35% gradient) where soils tend to be quite dry and sandy (i.e.
favorable conditions for brackenfem). B) Rice grass commonly grows in the
understory of flat landscape positions. Photos by T. Hobbs. ................................... 36

Figure 2.19- This kettle was plowed for tree planting. Rows of raised soil are still visible
on the ground (shown by black arrows). Only trees occupying backslope, shoulder,
and summit positions successfully grew. Photo by R. Schaetzl. ............................. 37

Figure 3.1- Schematic representation of a typical kettle in the study area (not to scale)
showing the locations of samples gathered within each kettle. ................................ 44

Figure 3.2- Location of sampled kettles on the Evart Upland. Kettles are outlined red. A
total of 53 kettles were sampled. Two samples were gathered at each site, resulting
in a total of 106 samples. .......................................................................................... 45

Figure 3.3- Particle size distribution curves for selection of backslope samples used in the
X-ray diffraction analysis. The range of grain sizes extracted from the samples is
20-53 pm. The lower limit of 20 um was chosen because backslope samples have
very little silt below that size. ................................................................................... 53

Figure 4.1- Aerial extent of silty sediments in the bottom of kettle site 17 (above). The
kettle area is shaded gray. The extent of silt is shaded brown. The red arrow gives
the perspective of the photograph at left. .................................................................. 59

Figure 4.2- Aerial extent of silty sediments in kettle bottom at kettle site 31 (above). The
area of the kettle is shaded gray. The extent of silt is shaded brown. The red arrow
gives the perspective of the photograph at left. ........................................................ 60

Figure 4.3- Particle size distribution curves of all 53 backslope samples. The average
particle size distribution is highlighted red. Blue and green curves represent the
coarsest and finest textured sample, respectively. .................................................... 63

Figure 4.4- Particle size distribution curves of all 53 kettle bottom deposit samples. The
average particle size distribution is highlighted red. Blue and green curves
represent the coarsest and finest textured sample, respectively. ............................... 64

Figure 45- Lower half of the USDA textural triangle, showing the texture of kettle
bottom samples (blue star) vs. kettle backslope samples (green x). Kettle bottom
deposit samples generally have loam and silt loam textures. Kettle backslope
samples are dominantly sand and loamy sand in texture. The main area of textural
overlap of kettle bottoms and backslopes is circled and shaded gray ....................... 65

xi

Figure 4.6- Average particle size distribution curves for all kettle bottom (red) and
backslope (blue) samples. Both populations have a bimodal particle size distribution
with peaks in the sand and silt range. The silt peak is different for each population
(18 um and 45 pm for bottoms and backslopes, respectively), while the sand peak is
the same (315 um). ................................................................................................... 68

Figure 4.7- Particle size distribution curves for three kettle bottom/backslope pairs that
have the greatest textural difference (i.e. have the greatest difference in silt content)
of the population of sample pairs. The kettle bottom/backslope pairs are color-
coded; solid lines represent kettle bottom samples, while the dashed lines represent
corresponding backslope samples. The colored bars at the top of the graph indicate
the difference in silt peak location between bottom/backslope pairs. ...................... 70

Figure 4.8- Particle size distribution curves for three kettle bottom/backslope pairs
representative of the average difference in silt content among the whole population.
The kettle bottom/backslope pairs are color-coded; solid lines represent kettle
bottom samples, while the dashed lines represent corresponding backslope samples.
The colored bars at the top of the graph indicate the difference in silt peak location
between bottom/backslope pairs. .............................................................................. 71

Figure 4.9- Particle size distribution curves for three kettle bottom/backslope pairs that
are the most texturally similar. The kettle bottom/backslope pairs are color-coded;
solid lines represent kettle bottom samples, while the dashed lines represent
corresponding backslope samples. The colored bars at the top of the graph indicate
the difference in siltpeak location between bottom/backslope pair. ........................ 73

Figure 4.10- Vectors showing the amount of downslope coarsening (blue) and fining
(red) with respect to mode sand grain size between backslopes and kettle bottoms.
Each vector contains two sets of x-y coordinates. The starting position of the arrow
is the sand peak in backslope samples. The end position of each arrow represents
the sand peak in the adjacent kettle bottom. Large changes in the x-coordinate
denote significant coarsening or fining downslope. Large changes in the y-
coordinate denote significant differences with respect to the amount of sand between
the two landscape positions. ..................................................................................... 78

Figure 4.11- Photograph looking downslope from the rim of kettle site 33. A gully is
visible in the foreground and leads to a fan-like deposit downslope in the kettle
bottom. Photo by T. Hobbs. ...................................................................................... 80

Figure 4.12- Photograph showing the top of the kettle bottom deposit profile at site 33.
The fan deposit shown in Fig. 4.11 is preserved as a sandy layer fi'om 0-5 cm in this
photo. This erosional event buried the charcoal rich soil from 5-15 cm. Photo by R.
Schaetzl. .................................................................................................................... 81

xii

Figure 4.13- Location of stable site-1 relative to the other stable site locations. B)
Topographic map showing surface characteristics of stable site-1. C) Photograph of
surface vegetation at this site. ................................................................................... 84

Figure 4.14- Depth plot showing distribution of four silt components with depth at stable
site—1 .......................................................................................................................... 85

Figure 4.15- A) location of stable site-2 relative to the other stable site locations. B)
Topographic map showing surface characteristics of stable site-2. C) Photograph of
surface vegetation at this site. ................................................................................... 86

Figure 4.16- Depthplot showing distribution of four silt components with depth at stable
site-2 .......................................................................................................................... 87

Figure 4.17 A) location of stable site-3 relative to the other stable site locations. B)
Topographic map showing surface characteristics of stable site-3. C) Photograph of
surface vegetation at this site. ................................................................................... 88

Figure 4.18- Depthplot showing distribution of four silt components with depth at stable
site-3 .......................................................................................................................... 89

Figure 4.19- Ordinary kriged grids of three silt components: coarse silt (left), medium silt
(middle), and fine silt (right) between backslope samples (top row) and kettle
bottom samples (bottom row). North is up in all of the grids. Black dots represent
sample locations across the study area. Greater silt content is represented by
increasingly darker color in both sample types. However, the values of percent silt
are different for each sample type because kettle bottoms have far more silt than
backslopes. ................................................................................................................ 94

Figure 4.20- A) Location of site 17 in relation to other depth sampled kettles on the Evart
Upland. B) Topographic map showing kettle site 17 with additional basin
characteristics (right). ............................................................................................... 98

Figure 4.21- Depositional sequence of kettle bottom sediments at site 17. The dashed
line marks the top of the coarse, sandy substrate sediments. The solid line marks the
top of a buried soil. ................................................................................................... 99

Figure 4.22- A) Location of site 24 in relation to other depth sampled kettles on the Evart
Upland. B) Topographic map showing kettle site 24 with additional basin
characteristics (right). ............................................................................................. 100

Figure 4.23- Depositional sequence of kettle bottom sediments at site 24. The dashed

line marks the top of the coarse, sandy substrate sediments. The solid line marks the
top of a buried soil. ................................................................................................. 101

xiii

Figure 4.24- A) Location of site 33 with in relation to other depth sampled kettles on the
Evart Upland. B) Topographic map showing kettle site 33 with additional basin
characteristics (right). ............................................................................................. 102

Figure 4.25- Depositional sequence of kettle bottom sediments at site 33. The dashed
line marks the top of the coarse, sandy substrate sediments. The solid lines mark the
top of buried soils .................................................................................................... 103

Figure 4.26- A) Location of site 42 with in relation to other depth sampled kettles on the
Evart Upland. B) Topographic map showing kettle site 42 with additional basin
characteristics (right). ............................................................................................. 104

Figure 4.27- Depositional sequence of kettle bottom sediments at site 42. The dashed
line marks the top of the coarse, sandy substrate sediments. The solid line marks the
top of a buried soil. ................................................................................................. 105

Figure 4.28- A) Location of site 45 with in relation to other depth sampled kettles on the
Evart Upland. B) Topographic map showing kettle site 45 with additional basin
characteristics (right). ............................................................................................. 106

Figure 4.29- Depositional sequence of kettle bottom sediments at site 45. The dashed
line marks the top of the coarse, sandy substrate sediments. The solid line marks the
top of a buried soil. ................................................................................................. 107

Figure 4.30- A) Location of site 49 with in relation to other depth sampled kettles on the
Evart Upland. B) Topographic map showing kettle site 49 with additional basin
characteristics (right). ............................................................................................. 108

Figure 4.31- Depositional sequence of kettle bottom sediments at site 49. The dashed
line marks the top of the coarse, sandy substrate sediments. The solid lines mark the
top of buried soils .................................................................................................... 109

Figure 4.32- A) Location of site 52 with in relation to other depth sampled kettles on the
Evart Upland. B) Topographic map showing kettle site 52 with additional basin
characteristics (right). ............................................................................................. 110

Figure 4.33- Depositional sequence of kettle bottom sediments at site 52. The dashed
line marks the top of the coarse, sandy substrate sediments. The solid line marks the
top of a buried soil. ................................................................................................. 111

Figure 4.34- A) Location of site 53 with in relation to other depth sampled kettles on the

Evart Upland. B) Topographic map showing kettle site 53 with additional basin
characteristics (right). ............................................................................................. 112

xiv

Figure 4.35- Depositional sequence of kettle bottom sediments at site 53. The dashed
line marks the top of the coarse, sandy substrate sediments. The solid lines mark the
top of buried soils .................................................................................................... 113

Figure 4.36- Simplified versions of the depthplots in figures 4.20 - 4.35, ranked in order
of increasing susceptibility to slopewash deposition from the surrounding
backslopes: steepest backslopes and smallest area-to-drainage basin ratio (left) to
shallowest backslopes and largest area-to-drainage basin ratio (right). These profile
representations are not drawn to scale relative to each other .................................. 116

Figure 4.37- This photograph shows a trough containing the sequence of kettle bottom
sediments at site 53. The pale tan sediments at the bottom of the photo are the sandy
substrate sediments (lowermost C-horizon). Overlying the buried soil is the first
interval of silty, aeolian sediment (brown colored). Photo by R. Schaetzl ............ 120

Figure 4.38- A) Photo and data from the soil profile excavated at site 33. This site
contains four buried soils, with charcoal randomly spread throughout the profile, but
mostly concentrated within the soils. B) Depth plot showing the down profile
textural variation at site 33. Black lines indicate the top of buried soils. C) Close up
of the lowermost buried soil from which charcoal was extracted at ~ 120 cm. Note:
the one-meter mark on the tape measure is above the extent of the picture, therefore,
the 20 cm mark in this frame is actually 120 cm. Photos by R. Schaetzl. ............. 121

Figure 4.39- A) and B) Woody charcoal pieces commonly found in buried soil at site 33.
C) Seeds were also found in the buried soil. D) Many small fragments of charcoal
sticks (~0.25-0.5 mm) within the matrix of sand grains. ........................................ 122

Figure 440- Pieces of amorphous, black-colored silica found interspersed within the
lowermost buried soil at Site 33 .............................................................................. 123

Figure 4.41 - Particle size distribution curves for two of the most texturally different kettle
bottom-backslope pairs. The range of silt-sized particles extracted from both is
outlined (20-53 microns) ......................................................................................... 129

Figure 4.42- Bottom: sites 24 and 42 both have distinct intervals of aeolian deposition in
the kettle bottom as indicated in the depthplots (intervals circled). Above: Particle
size distribution curves of the two aeolian intervals and their adjacent backslopes.
The 20-53 pm fraction of these samples and their adjacent kettle backslope samples
were extracted for X-ray diffraction analysis. ........................................................ 130

Figure 4.43- Results of the X-ray diffraction analysis for the two most texturally different
kettle bottom-backslope pairs. Each kettle bottom-backslope pair is color-coded.
Dashed lines represent backslope samples, while solid lines represent kettle bottom
samples. Each diffraction pattern is offset 500 counts above the underlying pattern
for ease of visual comparison. As such, the Y-axis should not be read as an absolute

XV

count. Three significant minerals are present in the samples: Quartz, K-feldspar,
and plagioclase feldspar. ......................................................................................... 131

Figure 4.44- Results of the X-ray diffraction analysis for the two kettle bottoms
containing distinct aeolian units. Each kettle bottom-backslope pair is color-coded.
Dashed lines represent backslope samples, while solid lines represent kettle bottom
samples. Each diffraction pattern is offset 500 counts above the underlying pattern
for ease of visual comparison. As such, the Y-axis should not be read as an absolute
count. Three significant minerals are present in the samples: Quartz, K-feldspar,
and plagioclase feldspar. ......................................................................................... 132

Figure 5.1- A Holocene record of aeolian activity from Elk Lake, MN. Black shading
represents % Na (higher values indicate increased dryness). Line pattern shading
indicates mass accumulation rate. The red bracket highlights a cycle of increased
aeolian activity that corresponds with the oldest radiocarbon date obtained in this
study. Diagram afier Dean (1997). ......................................................................... 142

Figure 5.2- Diagram showing the comparison of hydroclimate histories reconstructed
from Hole Bog, MN (Blue) and Minden Bog, MI (red) for the past 2500 yrs after
(Booth et al., 2006). Both records show a prominent dry interval around 1000 cal.
years BP (highlighted yellow). This time period corresponds to the radiocarbon date
of 920 cal. years BP obtained in this research. ....................................................... 144

xvi

CHAPTER 1 - Introduction

In many glaciated landscapes there are areas of hummocky topography containing
numerous internally drained depressions, called kettles (Bennett and Glasser, 1996).
Kettles are enclosed, bowl-like hollows in the landscape, marking the spot where former
detached masses of glacial ice melted in place. As topographically closed and internally-
draining depressions, kettles collect and retain sediment over time, thereby preserving a
record of sediment inputs in the bottom of the basin (Walker and Ruhe, 1968). The
sediments in kettles can, therefore, be studied to reconstruct the geomorphic history of
the basin. To that end, areas with many adjacent kettles may provide opportunities to
study the geomorphic history of the landscape they occupy.

This study focuses on a densely kettled upland (herein referred to as the Evart
Upland) in the northwestern Lower Peninsula of Michigan, near the town of Evart. The
Evart Upland lies within the southern portion of the North-Central interlobate drifi realm
of Michigan, between the former Lake Michigan and Saginaw lobes of the Laurentide ice
sheet (Rieck and Winters, 1993) (Fig. 1.1). The landforms and drift in this part of
Michigan are largely composed of sandy sediments associated with deposition via
glaciofluvial processes (Rieck and Winters, 1993; Schaetzl and Weisenborn, 2004;
Schaetzl and Forman, 2008). Situated in an interlobate reentrant, the Evart Upland likely
formed as glaciofluvial sediments buried a complex arrangement of stagnant ice. As
such, this landscape is characteristically sandy, and upon melting of the buried ice,

became topographically complex with many kettles.

 

  

 

Major Glacial Lobes and End MorainesJ

 

 

:‘A

@h‘ppewg’lobefi

 

flvhHuron-Erie
.lobe

 

   
 
 

- End moraines and large
heads of outwash

 

 

 

Figure 1.1- The Evart Upland is a complex assemblage of outwash with many kettles, situated
between the Lake Michigan and Saginaw lobes of the former Laurentide ice sheet. The Evart
Upland is in central Osceola County, MI (highlighted red). Images in this thesis are presented in
color.

Initial field investigations of the Evart Upland revealed that many of the kettles
there have a deposit of silty sediment in the lowest portion of the depression. These silty
deposits are ~lm thick, and transition abruptly to coarse textured, sandy outwash below.
Similarly, the surrounding kettle backslopes are largely composed of sandy drift.
Furthermore, some of the kettle bottoms on the Evart Upland contain buried soils within
the sequence of silty sediments. Figure 1.2 (below) is a schematic representation of a

typical kettle in the study area.

|
3.
§
5’ ‘
‘ a
. 7/ -
‘\ ' ' '
sf ,
' ' n u g
,1 ‘0 . o
D '
‘
i. J
r J ,1
i
.............
.................
....................
.......................
........................
............................
.............................
...............................
...............................
.................................
................................
..................................
...................................
.....................................
.....................................
.......................................
.......................................
.........................................
.........................................
...........................................
...........................................
.......................
. .

Sand dnfi ............ Silty sediments .......... Sand dnfi
..............
.................
...................
............................................................
.................................................................
........................................................................
.................................................................................
................................................................................
.................................................................................
................................................................................
.................................................................................
................................................................................
.................................................................................

 

 

 

Figure 1.2- Cross section of a typical kettle in the Evart Upland, containing silty sediments in the
lowest part of the depression. (not to scale)

Other researchers have documented similar deposits in kettle bottoms elsewhere,
specifically in the Grayling Fingers region of Michigan (Schaetzl and Weisenbom, 2004;

Schaetzl, 2008). Because deposits of silt in the kettle bottoms on the Evart Upland

appear to be anomalous with respect to the preponderance of sandy sediments that

dominate this region, the purpose of this research is to determine the likely geomorphic
origin(s) for the silt, and assess its paleoenvironmental importance.

There are two possible geomorphic scenarios by which silty sediments could be
deposited in the kettles on the Evart Upland. Both of these scenarios are outlined below

in the form of research hypotheses:

Hypothesis 1- the silt was winnowed from the surrounding upslope sediments

This hypothesis suggests that the surrounding backslopes were the source of silt in
kettle bottoms. A textural difference between the bottomlands and surrounding
backslopes is not uncommon in internally drained hillslopes, such as kettles. In fact, on
most steep, internally drained hillslopes, normal erosion and downslope sorting processes
result in a gradual sequence of progressively finer textured soils in the downslope
direction because smaller particles are often preferentially entrained, and more easily

transported further downslope (Milne, 1936; Walker and Ruhe, 1968; Malo et al., 1974).

Hypothesis 2- the silt is loess (aeolian silt)

This hypothesis suggests that the silt in kettle bottoms was brought into the kettles
by wind. Schaetzl and Weisenbom (2004) and Schaetzl (2008) concluded that the silty
deposits in the Grayling Fingers region of Michigan, which overlie sandy drift on stable,
flat uplands there (and in kettle bottoms on those same uplands) most likely have an
aeolian origin (i.e. loess). Given that the Grayling Fingers region is less than 150 km
northeast of the Evart Upland, it is possible that the silty sediments in kettle bottoms on

the Evart Upland have a similar origin.

The two hypotheses outlined above suggest drastically different
paleoenvironmental interpretations for the post-glacial period in this part of Michigan. A
slopewash origin for the kettle bottom silt suggests that precipitation in the past has been
significant and consistent enough to generate runoff, continually transporting sediments
downslope. This hypothesis also presumes a type of uniformitarianism— that the same
geomorphic processes (slopewash) observable on many hillslopes today, have been
operating in the Evart Upland since the landscape stabilized after deglaciation. In
contrast, an aeolian origin for the kettle bottom silt suggests that there were
exposed/unstable surfaces nearby with sufficient silt content, and at some time in the
past, there existed ample winds to transport that silt to the Evart Upland. Moreover, this
hypothesis suggests that different processes (other than what are observable today) were
operative in this landscape in the post-glacial period. These two hypotheses are tested in,
and form the core of, this thesis.

The Evart Upland is primarily state forest land, allowing for full, unrestricted
access to sample sites. The great abundance of kettles in the study area provided multiple
field sites to sample, thereby facilitating the creation of a robust and spatially extensive
data set. The primary field methods used here included extensive sampling of soils from
different landscape positions, (such as kettle bottoms, backslopes, and other
geomorphically stable sites), in order to characterize their texture and compare their
particle size distributions, both with depth at specific sites, and across the study area. The
primary lab method used here involved analyzing the textural properties of these samples
by laser particle size analysis. Other lab methods, such as radiocarbon analysis and X-ray

diffraction, were used to answer additional research questions (listed below). In addition,

publicly available geospatial data (such as, topography, soils, etc.) were used in
conjunction with textural data in a geographic information system to 1) support/refute the
two aforementioned hypotheses regarding the origin of the kettle bottom silt, and 2)
characterize the geomorphology of the Evart Upland.

A number of additional questions, prompted by preliminary analysis of the data,
were generated throughout the course of this research. These questions did not easily fit
into a hypothesis-testing framework, and were answered by gathering additional data
from the field. Data obtained via additional field sampling provided supplementary
evidence to support/refute hypotheses regarding the origin of the kettle bottom silt.

These questions are listed below:

1. How do kettle bottom sediments change (texturally) with depth?

2. When was silt deposited in kettle bottoms in the study area?

3. If the silt in kettle bottom is aeolian, is it local or from an extra regional source, or

both?

Details of the methods used to address these questions, and the two aforementioned
hypotheses (as well as their results), are presented in the following chapters, in the order
they are listed here.

The Evart Upland was chosen as a study area because relatively little is known
about the geomorphology of this part of the Lower Peninsula of Michigan. Therefore, the
abundance of kettles provided an opportunity to establish a base of knowledge about the
geomorphology of an area previously undocumented. The results of this research will
hopefully draw attention to the importance of dry kettles as repositories of geomorphic

information in formerly glaciated areas.

CHAPTER 2 - Study Area

W

The study area (the Evart Upland) is a 68 km2 area of sandy, rolling hills located
in Osceola County, MI. The upland is centered on, and encompassed by several USPLS
townships: Township 18N, Range 8W; Township 19N, Range 8W; Township 18 N,
Range 9W; and Township 19N, Range 9W, of Osceola County, MI. South of the upland
is the Muskegon River and the city of Evart. The Evart Upland is primarily forested, and
is largely owned and managed by the state of Michigan as part of the Pere—Marquette
State Forest. US. Highways 131, 10, and 115 pass to the west, south and northeast of the
study area, respectively (Fig. 2.1). Access to the Evart upland is possible via county and

seasonal roads, and various two-track paths.

Wm!

The extent of the Evart Upland is herein defined by a combination of topographic
and hydrologic characteristics. The Upland proper is characterized by highly undulating
and hummocky terrain, containing numerous dry kettles. The lowlands that surround the
Evart Upland are characterized by low relief, wet, hummocky terrain (i.e. most of the
kettles are bogs and wetlands, often containing ponded water). Indeed, I define the
boundary between the Evart Upland and the surrounding lowlands as the elevation where

dry upland kettles transition to wet lowland depressions. This boundary, however, is

 

 

 

0 2.5 5 10 15 20

(-

Elevation N

Figure 2.1- Color, hill-shaded, digital elevation model (DEM) of Osceola County, MI. The extent of
the Evart Upland is shaded gray. The dashed line represents the Pare-Marquette State Forest
boundary. The black lines are major highways.

595 meters _

170 meters

 

 

 

 

more appropriately envisioned as a transitional zone, since there is no single elevation
contour that consistently separates the dry upland landscape from the wetter lowlands. In
actuality, there exists a range of elevations where closed depressions transition from dry
to wet, around the perimeter of the upland. In the southwestern portion of the study area,
kettles transition from dry to wet between 370-390 meters elevation (Fig. 2.2). In the
northeastern portion of the study area, kettles transition from dry to wet between 390-410

meters elevation (Fig 2.3).

mm

The soils and sediments of the Evart Upland are generally very dry, having been
formed mainly in well-drained, sandy sediment. As such, the sediments have deep water
tables, and therefore, few depressions within the upland contain permanent wetlands. As
noted above, however, bogs and wetlands are very common in the transitional zone
immediately surrounding the upland, where groundwater discharges into hummocky
closed depressions. The surrounding lowlands are, therefore, characterized by numerous
bogs, wetlands, and lakes separated by low relief (3-6m) hummocks (Fig. 2.3). Some
significant lakes in the surrounding lowlands include Silver Lake, Strawberry Lake, and
Hicks Lake. Most of the lakes in the surrounding lowlands are kettle lakes. However,
Little Long Lake and Big Long Lake occupy narrow curving valleys, suggesting that they

could be of subglacial (meltwater erosional channel) origin (Fig. 2.4).

 

 

0 0.25 0.5 1 1.5 2

Figure 2.2- Combined topographic map (3-meter contour interval) and color-shaded DEM of the
western portion of the Evart Upland. The black line separates dry kettles on the upland from bogs,
wetlands, and lakes in the lowlands, and is the approximate boundary of the study area.

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2.3- Combined topographic map (3-meter contour interval) and color-shaded DEM of the
northeastern portion of the Evart Upland. Lower relief, wet, hummocky terrain characterizes the
surrounding lowlands to the northeast.

 

Elevation

0 1.5 3 6 9 12_ 520'“

   

285 m

Figure 2.4- Grayscale DEM of the Evart Upland and surrounding areas, showing hydrologic
features. Blue lines represent perennial streams and rivers (data source: Osceola County
Hydrography shapefiles from the Michigan Geographic Data Library (Michigan Department of
Information Technology, 2007)). The Muskegon River, located south of the upland, flows to the
southwest. The Hersey River, flowing south, joins the Muskegon River at the bottom of the map.

In general, most perennial streams in and around the study area flow radially
away fiom the Evart Upland, eventually making their way to larger rivers. Major rivers
in the area include the Muskegon and the Hersey, which border the Evart Upland to the
south and west, respectively. A network of weakly integrated stream valleys extends up
the central portion of the upland from the south and southeast. These stream valleys
subtly transition upslope to dry valleys containing kettle chains (discussed below). Twin

Creek and its ephemeral tributaries are the most notable of such features.

2.4 Geomorphology

The Evart Upland is situated between the former Lake Michigan and Saginaw
lobes of the Laurentide ice sheet (see Fig. 1.1, Introduction chapter). As is typical of
interlobate terrain (Rieck, 1976), the topographic expression of the upland is undulating
and complex. This complexity can likely be attributed to the wastage of detached
stagnant ice masses, which resulted in a number of localized closed depressions (kettles).
Because this study emphasizes kettles as repositories of geomorphic and sedimentologic
information, the majority of the geomorphic characterization below focuses on the

distribution and morphology of the kettles themselves.

2.4.1 Methods Used to Characterize the Geomorphology of the Study Area

The geomorphic interpretations made herein are largely based on the analysis of
digital USGS topographic maps for the study area. As such, most of the interpretations

are not verified by analysis of the landforms in the field. Where applicable, references to

13

supporting literature are made regarding the formation of similar landforms in nearby
regions. This characterization was done to assist in analysis of the field data and
discussion of the landscape.

The primary method of characterizing this landscape involved mapping the
locations of closed depressions (kettles). The purpose of mapping kettles was to better
understand the surface morphology of the upland and to aid in selecting sample sites.
Major kettles within the boundary of the Evart Upland were digitally mapped in a GIS,
using ArcMap 9.2 (ESRI, Redlands, CA). Kettles too small to be identified on USGS 7.5
minute topographic quadrangles (i.e. shallower than the three meter contour interval)
were not mapped.

A Digital Raster Graphic (DRG) of USGS topographic quadrangles for Osceola
County was used as the data base for the kettle mapping exercise. The rims of kettles
(i.e., their perimeters), defined as the highest closed contour line of each depression, were
individually traced on the DRG in ArcMap. The editing was done at a scale of 123,500 to
ensure that the perimeters of kettles were accurately traced. Kettle depth was recorded
during the mapping exercise by determining the elevation difference between the deepest
part of the depression (as best estimated) and the kettle rim. The mapped kettles were
projected in Michigan Georef (meters) so that the plan geometry (perimeter and area) of
each kettle could be calculated. All other spatial data depicted in the maps in this chapter
were gathered from the Michigan Geographic Data Library (Michigan Department of

Information Technology, 2007).

14

2.4.2 Topographic Characteristics

The surface expression of the Evart Upland is undulating, very hummocky, and
complex, with little or no external drainage. Four observations are noteworthy, with
regard to the topography of the upland: 1) kettles are very common, 2) massive sandy
ridges, devoid of kettles, are also common, 3) in the low areas between the ridges, many
kettles (often linked in chains) exist, and 4) the ridges and kettle chains exhibit a
conspicuous linearity, dominantly trending NW-SE. Each of these topographic traits is
discussed individually in the sections that follow, along with some possible
interpretations of their origin.

The irregular and dome-like Evart Upland rises up to 150 m higher than the
surrounding lowlands (Fig. 2.5). Local relief on the upland (between kettle bottoms and
adjacent summits) can be up to 24 m, but is typically about six meters. The highest
hilltops in the central portion of the Evart Upland peak at about 460-470m above sea
level (asl). Relief between the Upland and the Muskegon River floodplain to the south is
about 170 m (Fig. 2.5).

The distinct ridge that trends north/northeast, radiating out, away from the east
side of the Evart Upland (Fig. 2.5), has a subdued surface expression relative to the
higher, central portion of the upland. Topographically, this ridge is similar to the
southeastern portion of the upland, where relief between closed depressions and adjacent
summits is generally about 3-6 m (i.e. the kettles there are generally small). Auger
shavings from road cuts through the ridge suggest that it is comprised of sandy loam drift,

making it texturally different than the rest of the Evart Upland, which is dominated by

15

 

0 2.5 5

10
"ilometers

[__ __ -_

Topographic Profile

 

 

 

  
    

....................... L5ii’fe’liéf
..... hummocks

so ................................... .l

skegon River

 

 

. 0 5,000 10 .000 15,000 20 .000

A——H0rizontal Distance (meters)———B ‘

Figure 2.5- A) Grayscale DEM of the Evart Upland showing the location of transect line A-B (red)
that was used to construct the topographic profile shown in part B. The NNE trending ridge
extending from the eastern portion of the study area is not considered part of the Evart Upland. B)
Topographic profile of transect line A-B. Kettles on the upland have much greater local relief than
kettles in the surrounding lowlands near the Muskegon River.

sand-textured outwash. It is likely that this ridge is an end moraine of the Saginaw lobe,
and thus, not an interlobate feature at all. Therefore, in this study, the ridge is not
considered part of the Evart Upland. As such, it was not investigated with the same level

of detail as the remainder of the upland.

2.4.3 Kettles: Range of Morphometric Characteristics

A total of 535 major kettles were identified and mapped within the extent of the
study area, for a density of nearly eight major kettles per km” (Fig. 2.6). Kettles on the
Evart Upland exhibit a wide range of morphometric characteristics such as, depth, area,
perimeter, and perimeter to area ratio (plan view shape) (Table 2.1). Some kettles are a
single depression; others are composite, having more than one depression encircled by
the highest closed contour line. There are also differences in kettle density and size with
respect to location on the upland. On the northwestern and central portions of the upland,
larger, more widely spaced kettles are common. On the southeastern portion of the

upland, kettles are smaller and more closely spaced (Fig. 2.7).

Table 2.1- Morphometric characteristics of kettles.

 

Depth (m) Area (sq. m) Perimeter (m) Area to perimeter

 

ratio

Mean 6 9,670 337 0.08
Max 24 160,000 2,455 0.23
Min <3 260 62 0.01

 

 

' There are many small kettles on the ridge that extends to the north/northeast that were not mapped.

l7

 

Elevation

0 0,5 1 2 3 4 , 520 meters
km '

   

 

 

285 meters

Figure 2.6- Grayscale DEM of the study area with major kettles highlighted in red. A total of 535
kettles were mapped within the boundary of the Evart Upland.

" if

‘ \ , r .‘ _ .u ‘4 ‘7‘.“ K ' ‘
\ -' ' .4 '%%-§ ,rfi
'0 1’: “Tr-:6. l’..\|‘/hgg \ ",1.

 

 

 

 

Writs“

Figure 2.7- Grayscale DEM showing two areas on the Evart Upland with distinct kettle morphology
(below); one from the western portion (upper left), the other from the southeastern portion (upper
right). The two topographic maps above (ll-meter contour interval) are the same scale.

2.4.4 Ridges

There are many distinct, elongate and round, conical (convex upward), mound-
like landforms on the Evart Upland. For descriptive purposes, these landforms will
herein be referred to as ridges (implying no genetic origin) because their sedimentology
and internal structure were not explored in the field. Some of the ridges in the study area
are broad and roughly linear. The most conspicuous of the ridges are massive and semi-
conical, containing a number of localized peaks and gullied sides. The highest ridges
peak at about 470 m in elevation. Some ridges are small and linear, while others are
steep-sided and relatively flat-topped (2.8). Kettles are absent on nearly all of the ridges,
but almost always surround them on the sides. Local relief between the ridge tops and
surrounding kettles can be up to 70 meters.

In plan view, these ridges somewhat resemble the hammocks and hummock tracts
described by Ham and Attig (1996) and Johnson et a1. (1995) for stagnant ice landscapes
in Wisconsin. However, unlike the hummocks they described, the ridges in the Evart
Upland are interspersed with dry kettles (both kettle chains and groups of kettles), rather
than ice-walled lake plains and wetlands, and therefore, likely have a different
geomorphic origin. The abundance of dry kettles and ubiquity of well-sorted, sandy
sediments across the study area suggest that the ridges are the result of glaciofluvial
deposition, either within large crevasses in the former stagnant ice, or around isolated
masses of stagnant ice. Upon full meltout of the stagnant ice and stabilization of the
landscape, these sandy infillings then became positive topographic features. Similar
interpretations were made for the formation of kettled portions of the Lake Michigan/

Saginaw interlobate terrain further south near Hastings, MI (Folsom, 1971; Rieck, 1979).

20

 

Figure 2.8- USGS topographic map (3-meter contour interval) showing examples of ridges (ridge
tops shaded tan) next to kettles (shaded red) in the study area. The ridges in the left map are
relatively flat topped and have steep sides. The ridge in the right map is massive with deeply gullied
sides.

2.4.5 Low Areas Between Ridges

Between the largest ridges on the Evart Upland, there are low areas with many
kettles (Fig. 2.9). These areas range in size, but are generally about 0.5 to 1.0 km2 in
area. Kettles within these low areas are sometimes linked in chains or are elongated
parallel to each other; in other areas they are more randomly oriented. Some of the low
areas contain groups of many adjacent kettles, and are surrounded by ridges that have
multiple small peaks. The ridges surrounding low areas sometimes have steep sides,
indicating that they were abutted by ice during their formation, i.e., they are ice-contact
landforms. As such, it may be that the low areas containing many kettles were once filled
to at least the elevation of the surrounding steep ridges with stagnant ice, and not until
later did the much larger ice mass disintegrate into individual ice masses, eventually

forming groups of kettles in topographically lower areas.

21

 

Figure 2.9- Combined USGS topographic map (3-meter contour interval) and color-shaded DEM
showing low areas (yellow shading) between ridges (brown shading). In these low areas kettles are
grouped. Major ridge crests are traced with black lines. Kettles are shaded red.

22

2.4.6 Kettle Chains

Some of the low areas between ridges on the Evart Upland contain kettle chains.
The individual kettles comprising these kettle chains gradually become more integrated
in the downstream direction, eventually transitioning to intermittent, and then perennial,
stream valleys downslope (Fig 2.10). For example, the low area between the ridges in
sections 1, 2, 11, and 12 of T8N, R9W contains a network of kettles that subtly transition
into a series of ponds (i.e., Whitmore pond on the topographic map) that are then drained
by Twin Creek further downstream (Fig. 2.11). Kettles that lie just beyond the highest
reach of this drainage network are topographically isolated (not integrated by stream
incision). The long axis of these non-integrated kettles is usually oriented parallel to the
encroaching stream valley. Within the middle reaches of the valley, kettles are somewhat
integrated; the highest closed contour line of adjacent kettles is usually the same
elevation. However, in the lower valley reaches there are no dry kettles. Instead, ground
water emerges to form a series of deranged shallow ponds and wetlands, which are
drained by Twin Creek.

Schaetzl and Weisenbom (2004) have interpreted kettle chains in the Grayling
Fingers region of Michigan (150 km northeast of the study area) as palimpsest features
that inherited their orientation as ice overrode, and was preferentially preserved in,
preexisting fluvial valleys. Their interpretation is supported by topographic evidence
showing that 1) the kettle chains are sometimes arranged in a meandering pattern, yet 2)
there are ridges that separate individual kettles within the chains. Therefore, it was
suggested that their arrangement could not be the result of post-glacial stream incision

(Schaetzl and Weisenbom, 2004). In the Evart Upland, however, the rims of adjacent

23

kettles in the middle and lower valley reaches are usually similar in elevation (not
separated by significant ridges). Furthermore, the long-axis orientation of the kettles is
usually parallel with both the valley they occupy and the perennial streams that drain the
upland, neither of which meander significantly. Therefore, the stagnant ice masses that
eventually formed the kettles may have been sources of meltwater in the late phases of
deglaciation, facilitating development of the low valleys they occupied. This
interpretation suggests that the perennial streams that drain the upland inherited their

drainage pattern from the distribution of former ice masses, probably in linear chains, that

occupied low areas between ridges on the Evart Upland.

 

Figure 2.10— Combined hillshade and color-shaded DEM presented in a three-dimensional view
looking northwest, at the Evart Upland. The boundary of the upland is outlined in black. The red
lines trace weakly integrated stream valleys that are headed by kettle chains upslope.

24

 

 

 

 

0
1‘ y(o
W

 

.1 9
g
(10‘ o
s 0‘
~.\
‘ W
. C
‘L-- ('4' '- D e
1\
3 D
‘ V 0‘
0 \;“?:n\
is] r Ni ‘.V
‘4 ’5',‘ K r l
Jcégékikx‘ .‘
V\
i k." m“ 4|
0 ' \,

 

TWin Creek

.V all

0 250 500 1 .000 1 .500 2.000

Figure 2.11- A) Topographic map (3-meter contour interval) of a portion of the Evart Upland. The
red polygons outline kettles. The blue line outlines the drainage basin of kettles that subtly transition
to Twin Creek, south of the map. Note that the long axis orientation of the kettles that occupy the
central portion of the valley is more or less parallel to the valley. Kettles in the northern most
portion of the drainage basin are not well integrated. B) Inset map of a grayscale DEM showing the
location of the drainage basin (shaded blue) on topographic map in (A). The Evart Upland is
outlined in black. Blue lines on the inset map are perennial streams.

25

2.4.7 Linearity of Kettles and Ridges

Both the kettles and ridges described above exhibit a consistent linearity across
the study area, generally trending northwest-southeast. Superimposed on this trend is a
secondary, less consistent northeast-southwest trend, evident in some ridges and
depressions. Sometimes a single ridge or kettle possesses both trends. Even some of the
surrounding lakes, such as Silver and Horseshoe lakes, which lie just west of the upland,
possess both trends. This linearity manifests itself on multiple scales, from small kettles
on the upland to large lakes in the surrounding lowlands (Fig. 2.12). The full geomorphic
relevance of these observations is still unknown. However, it may be that the orientation
of these landforms was controlled by the structure of the former stagnant ice, which, in
this region was ultimately the product of two conjoining ice margins: the Lake Michigan
ice from the west and northwest, and the Saginaw lobe ice from the east and northeast.
Perhaps, upon meltout, the complex structural arrangement of stagnant ice then
influenced the path of glaciofluvial meltwater, thereby playing an important role in the
orientation of subsequent landforms.

In summary, the sandy sediments that make up ridges on the Evart Upland were
likely areas of ice-contact glaciofluvial deposition, either within fractures in the stagnant
ice, or as topographic lows on the former, ablating ice sheet surface. Evidence for this
interpretation includes the frequent occurrence of high ridges that usually surround low
areas containing many kettles. Some of these ridges have steep slopes, suggesting that
they are ice-contact landforms (Bennett and Glasser, 1996). Most of these landforms

have consistent long axis trends (NW-SE, and to a lesser extent, NE-SW), the orientation

26

 

Figure 2.12- Color shaded DEM showing examples of a composite kettle (A) and lake (B) that have
contrasting long axis trends as described in the text. Kettles are shaded red. Lakes are not shaded.

of which may have been inherited from the structure of the former stagnant ice, or its
margins. Although the length of time that stagnant ice persisted in the study area is
unclear, the presence of kettle chains within stream valleys suggests that ice masses
persisted within the valleys throughout a later phase of landscape erosion, perhaps even

until subaerial streams (still present today) became established.

M

Pedologic diversity across the Evart Upland is low. A total of five distinct, sandy
soil series together make up more than 97% of the study area‘. Table 2.1 lists these
series in order of abundance within the study area. The five most common soil series

mapped in the study area (in order of abundance) are Montcalm (coarse-loamy, mixed,

 

' This calculation was made by dividing the cumulative area of the five most common soil series by the
area of the Evart Upland boundary.

27

Table 2.2- Characterization data for the five most common soil series mapped on the Evart Upland ',
listed in order of abundance from most (top) to least (bottom).

 

 
   

Horizons2 Depth (cm) Texture Other Characteristics

          

lOYR 4/2

Ap ’ I T 0-1 loysd I Dig clszwell

B 18-23 lOYR 7/2 loamy sand ——P———'R“noff_ otem‘a}; 10W

Bhs 23-66 7.5YR 4/4 loamy sand 'Compnses 39'7 /° .°f Study area

' -Most common series, generally more

E 66'“ lOYR 6” loamy sand common along the lower portions of

Bt 81-91 5 YR 4/4 sandy loam the upland
E/Bt 91-152 lOYR 6/2 sand

loam san ,

 

 

 

 

 

 

 
   

  
  

    

c 152 + lOYR 6/3 p

..H .7-

   

   

10YR3/2 I sand -Drainag e calss: somewhat excessively

 

 

 

 

 

 

 

Ap 0-25
33 25-71 lOYR 4/4 sand -Runoff potential: low
E 71-114 lOYR 6/4 sand -Comprises 25.1% of study area
E/Bt 114 178 lOYR 6 / 4 d -Common on the highest, central
C 17;; + lOYR 6 /3 sand portion of the Evart Upland and
san .

       

0-3 -

 

 

O
A 3-8 lOYR 2/1 sand -Runoff DOICMZ N/ A
E 8-18 lOYR 6 /2 sand -Comprises 18.4% of study area . .
-Common rn lower landscape posrtlons
Bhs 18’3 8 7'5YR 4/4 sand such as deep kettles, valleys, and low
BS 38'69 7.5YR 5/6 sand areas between ridgCS.
BC 69-79 lOYR 6/3 sand

 

 

 

 

 

lOYR 6/3 sand

  
 

C 79-153

      

  
 

  
 

Ap I 10R 32/ ’ -Dralnage class: moderately well

 

sandy loam . '
Bs 15-51 lOYR 4/4 sandy loam -Run0ffmtem@l= high
- . o
Em 51-71 lOYR 5/2 loamy sand C°mpnses 12'4 /° °f Study area .
/ 1 -Common on the east/southeast srde of
E/Bt 71-91 lOYR 6 3 oamy sand the study area
Bw 91-132 7.5YR 4/4 sandy clay loam

 

 

 

 

 

 

 
   

132-152 7.5YR4/4 sand loam

    

O I 0-3 M h I i i - ‘ -Drainage lass: excessively
A 3-3 lOYR 3/2 sand -Runoff potential: very low
-Comprises 1.5% of study area

 

 

 

 

 

 

le 8-22 lOYR 5/4 sand . .
-Least common series, typically found
Bs2 ”'48 lOYR 5/6 sand in deeper kettles in the northwestern
BC 48"“ lOYR 6/8 sand portion of the study area
C 61-152 lOYR 6/3 sand

 

 

’ These data were compiled from a combination of SSURGO digital soils data (Soil Survey Staff, 2007)
and the Soil Survey for Osceola County, MI (Mettert, 1969).
2 Horizon designations used in Mettert (1969) were converted to current horizon nomenclature.

28

semiactive, frigid Alfie Haplorthods), Chelsea (sandy, mixed, mesic Lamellic
Udipsamments), Rubicon (sandy, mixed, frigid Entic Haplorthods), McBride (coarse-
loamy, mixed, semiactive, frigid Alfic Fragiorthods) and Grayling (sandy, isotic, frigid
Typic Udipsamments). Most of the soils on the Evart Upland formed in sandy parent
materials. Rubicon and Grayling soils are the sandiest of the five most common series on
the Evart Upland. Chelsea soils are slightly finer textured than Rubicon and Grayling.
Montcalm and McBride soils are the loamiest of the five most common series on the
upland, with McBride slightly finer textured overall than Montcalm. Therefore, in order
from coarsest to finest texture overall, the five most abundant series on the upland are:
Grayling -) Rubicon -) Chelsea '9 Montcalm-) McBride (Table 2.1).

Grayling and Rubicon soils are often mapped in lower landscape positions, such
as narrow valleys, deep, dry kettles, and low areas between ridges (Fig. 2.13). Rubicon
soils are abundant in the northwestern portion of the study area, and decrease in
abundance toward the southeast (Fig. 2.13). Chelsea soils are mostly mapped on the
highest, central portion of the Evart Upland and occasionally upslope of Rubicon series.
Like Rubicon soils, Chelsea soils decrease in abundance toward the east side of the study
area. Montcalm soils are the most abundantly mapped series and, in general, are evenly
distributed throughout the study area. Finally, McBride soils are common on the east and
southeastern side of the study area (Fig. 2.13).

The ubiquity of sandy parent materials on the Evart Upland is the greatest factor
(i.e. of the five soil forming factors outlined by Jenny (1941)) influencing the pedogenic

development of these soil series. In that context, precipitation (specifically, snowmelt) is

29

 

Soil Series
N Grayling

o 1.25 2.5 5 Che'sea

km - Rubicon
I _ a Montcalm
- McBride

Figure 2.13- Map of the five most common soil series on the Evart Upland (SSURGO digital soils
data (Soil Survey Staff, 2007)). All other mapped series are left blank (gray areas on the upland).
Grayling and Rubicon soils are mapped in narrow valleys, deep, dry kettles, and low areas between
ridges. McBride soils increases in abundance on the east side of the upland, while Rubicon soils are
generally absent there.

 

30

also important. Soils here likely do not freeze deeply due to a combination of their
location (just south of the mesic-fiigid soil temperature boundary (Schaetzl et al., 2005)),
and the insulating effect of thick snowpacks that exist here (Isard and Schaetzl, 1998).
Consequently, these soils may receive more infiltration from snowmelt in the winter than
soils fiuther north or west. As such, the combination of a cool climate, sandy parent
materials, and reasonably large amounts of snowmelt infiltration has led these soils to
develop many pedogenic characteristics of Spodosols (Schaetzl, 1996).

In summary, the five most abundant soil series on the Evart Upland exhibit a
range of pedogenic characteristics (Table 2.2), the development of which is mostly
influenced by the texture of their respective parent materials. In general, Grayling,
Chelsea, and Rubicon soils are sandy and excessively to somewhat excessively drained,
whereas Montcalm and McBride soils are loamier. Furthermore, the sandier soils are
usually mapped in lower landscape positions on the upland such as narrow valleys, deep,
dry kettles, and the low areas occupied by weakly integrated kettle chains described
above. In contrast, loamier soils are usually mapped on more stable landscape positions

such as the central high portion and the southeastern portion of the upland (Fig. 2.13).

2.6 Vegetation

2.6.1 Past and Present

According to the regional landscape ecosystem classification of Albert et a1.
(1995), the Evart Upland lies within the Cadillac sub-subsection VII.2.1. Vegetation
cover in this sub-subsection is characterized by northern hardwood forest species. At the

time of European settlement, the excessively drained sandy ridges in this area were most

31

likely covered by oak-pine forests containing Red pine (Pinus resinosa), White pine
(Pinus strobus), Red oak (Quercus rubra) White oak (Quercus alba), Black oak (Quercus
vilutina), Red maple (Acer rubrum), and Bigtooth aspen (Populus grandidentata) (Albert,
1995).

Presettlement vegetation cover on the Evart Upland was much different than in
the surrounding, wetter lowlands. According to Comer’s (1998) presettlement vegetation
map, the Evart Upland was primarily occupied by white pine-mixed hardwood forest
prior to European settlement. In contrast, beech-sugar maple-hernlock forests dominated
the surrounding lowlands. In general, the boundary of the upland coincides with the
presettlement vegetation boundary between white pine-mixed hardwood forest and
beech-sugar maple-hemlock forest (Fig. 2.14), as would be expected given that the
upland soils are xeric and the surrounding lowlands contain permanent lakes and
wetlands, set amidst generally loamier soils.

Today, forests in and around the Evart Upland are primarily reserved for timber
production. Significant portions of the study area lack canopy cover due to recent clear
cutting (Fig. 2.15). As such, forest cover on the upland is sometimes patchy, resembling
a mosaic of mixed hardwood forest in various stages of regrowth. Common species on
the upland are white pine, red pine, red maple, and bigtooth aspen. Other common trees

include: beech (F agus grandifolia), birch (Betula papyrifera), and oak (Quercus spp.).

2.6.2 Vegetation in Kettles

Some kettles on the Evart Upland are non-forested, allowing other vegetation
types, (mostly grasses and ferns) to emerge (Fig. 2.16). This pattern is easily identifiable

on the topographic maps of the study area (where kettles are shaded white), and on aerial

32

 

Presettlement Vegetation 0 0.5 1 2 3
7} Beech-Sugar Maple-Hemlock Forest

km
- Undifferentiated Swamp
‘ White Pine-Mixed Hardwood Forest

E Lake/River

Figure 2.14 Combined hillshade and presettlement vegetation map showing the Evart Upland and
surrounding areas. The study area boundary is outlined black. Presettlement vegetation data from
(Comer et al., 1998).

 

33

a.

i I
I
In!

' /

t

'-
V."—
_ ,. ,

".

 

-i\_

Figure 2.15- The Evart Upland is primarily used for timber production. Clear cutting is common.
Photo by R. Schaetzl.

T 9‘

 

Figure 2.16- Some kettles lack canopy cover. In the photograph above, the surrounding forested
summits open up to sweet fern and grasses (visible in the foreground) in the kettle bottom. Photo by
R. Schaetzl.

34

(.3 : ‘ I
10:3" 1 \§
.3 {7)

'.'

 

Figure 2.17— Topographic map (3-meter contour interval) of the northwestern portion of the study
area. Non-forested kettles are shaded white on this topographic map. Black arrows point to
examples of kettles that are non-forested. Other larger areas with white shading were clear-cut at
the time the topographic map was made (1983).

photographs (Fig. 2.17). Typically, the vegetation across a kettle changes gradationally
from summit to toeslope: forested summits gradually transition to toeslopes of grasses
and ferns. The most common fern in the study area is bracken fern (Pteridium
aquilinum). In Michigan, bracken fern mainly grows on dry, sandy soils (Schaetzl,
personal communication). Bracken fern is common on the steepest slopes on the Evart
Upland. Another common fern in the study area is sweet fern (Comptonia peregrina).
Commonly found growing next to bracken fern on the upland, sweet fern also favors dry,
sandy soils (NRCS, 2002). Finally, rice grass (Oryzopsis asperifolia) commonly grows

in the understory of some of the flatter landscape positions on the upland (Fig. 2.18).

35

 

Figure 2.18— A) Bracken fern is common on steep slopes (above photo looking up a kettle backslope
with a 35% gradient) where soils tend to be quite dry and sandy (i.e. favorable conditions for
brackenfern). B) Rice grass commonly grows in the understory of flat landscape positions. Photos
by T. Hobbs.

36

 

vtk‘

   

'5 1‘: H‘ -.'h‘ '9 -- ,'
. . _ . . . ... ..-_,.l:1.3;‘~'?~:'-tlw 5.. .. flat is. _. -_ -. l J
Figure 2.19— This kettle was plowed for tree planting. Rows of raised soil are still visible on the
ground (shown by black arrows). Only trees occupying backslope, shoulder, and summit positions
successfully grew. Photo by R. Schaetzl.

   

va- 5'“! ,h ..

_ 9'

In some kettle bottoms, the subtle linear microtopography left behind by
moldboard plows (evidence that the site was prepared for tree planting, probably in the
1930’s and 1940’s) is still visible along the ground (Fig. 2.19). Interestingly, only the
trees occupying backslope, shoulder, and summit positions have survived in some of
these kettles. The failure of re-growth in some kettle bottoms suggests that the
gradational change in vegetation cover from summit to kettle bottom is natural and may
depend on kettle shape and slope. The GLO (Government Land Office) survey reports of
the sections in the study area, compiled when this part of Michigan was first surveyed in
the late 1840’s, do not make reference to an absence of trees in kettle depressions.

However, in counties firrther northeast (sub-subsection VII.2.2, Grayling Outwash Plain),

37

it appears that GLO surveyors did note “large frost pockets” in depressions on the
outwash plain often containing dry prairie openings (Albert, 1995), suggesting that the
absence of forest vegetation in kettles here may be due to abnormally abundant frost

events, leading to overall shorter growing seasons.

2.7 Climate

In general, the climate in the study area is typical of humid continental-mild
summer locations. The climate station nearest the Evart Upland is about five km
southeast of the upland, in Evart. Table 2.3 lists monthly average maximum and
minimum temperatures, precipitation, and snowfall for station 202671 (Evart, MI) from
1971-2000. Given that the central portion of the Evart Upland is about 150 m higher and
five km north of the nearest weather station, average temperature and precipitation values
there may be slightly different from the weather station. The growing season length for
Evart, MI is about 114 days (starts 5/25, ends 9/ 15). Climate data presented in this
paragraph and in Table 2.3 were gathered from the NOAA Midwestern Regional Climate

Center (NOAA, 2008).

38

Table 2.3- Average monthly climate data for 30-year period (1971-2000).
Station: (202671) EVART, MI"

 

 

 

 

 

 

 

 

 

 

Month Max Temp. (°F) Min Temp. GEL Precipitation (in) Snowfall (in
Jan 26.8 8.3 1.9 16.3
Feb 30 8.8 1.5 1 1.4
Mar 40.3 19.2 2.4 7.9
Apr 54.6 31.3 2.8 1.5
May 68.6 41.3 2.9 0.0
Jun 76.5 49.8 3.2 0.0
Jul 81.8 55 2.5 0.0
Aug 79 53.2 4.1 0.0
Sep 70 45.2 4.6 0.0
Oct 57.6 34.3 3.0 0.2
Nov 43.9 26.7 2.3 5.0
Dec 31.4 15.7 2.1 12.3

[ Averafil 55 I 32.4 I 33.2 i 54.6 ]

 

' Data provided by the NOAA Midwestern Regional Climate Center, a part of the Illinois State Water
Survey at the University of Illinois at Urbana-Champaign under the Institute of Natural Resource
Sustainability, and located on the web at http://mrcc.isws.illinois.edu.

39

CHAPTER 3 - Methods

3.1 Preliminagy Work

3.1.1 Data Sources

An important prerequisite to the field component of this research involved
gathering the appropriate spatial data related to the physical features of the Evart Upland
and surrounding areas (i.e. topography, soils, hydrography, etc.). These data were
gathered to gain an understanding of the processes responsible for the formation of the
landscape, and to aid in hypothesis generation about the origin of the silty deposits in
kettle bottoms. To that end, topographic quadrangles (7.5 minute series) of the study area
were obtained from the United States Geological Survey (USGS). These paper maps
were pieced together and laminated for use in the field, both for keeping track of which
kettle sites were visited and occasionally to navigate the terrain on foot. Also, the
appropriate GIS shapefiles of Osceola County, M1 were downloaded from the Michigan
Geographic Data Library (Michigan Department of Information Technology, 2007).
These files included: public land units, roads, a digital elevation model (DEM), a digital
raster graphic (DRG) of the topographic quadrangles, NRCS-SSURGO soils data, and
hydrography. These shapefiles were added to an Argonaut TFlex laptop equipped with

ArcMap software (Redlands, CA) to aid in field navigation and analysis of the data.

3.1.2 Selection of Sample Sites

Initial, reconnaissance field investigations revealed that, more often than not, non-

forested kettles tended to have silty bottom sediments, whereas forested kettles tended to

40

have sandier bottom sediments. The details of the relationship between presence/absence
of forest cover and sediment texture in kettle bottoms is not fully understood, and
remains untested in this thesis. Nevertheless, non-forested kettles were preferentially
chosen as target sample sites because initial field investigation suggested (albeit
anecdotally) that they tended to have siltier kettle bottom sediments than forested kettles.
From the 535 major kettles mapped within the study area, a total of 159 target kettle sites
were identified and marked as a “target kettle” on a new point shapefile in a GIS. Then,

this file was loaded onto the laptop, and used as a guide to navigate to target sample sites.

3.2 Field Methods
3.2.1 Navigation

A 4-wheel drive vehicle, GPS unit, and laptop computer equipped with ArcMap
software, were used in combination to navigate to target kettle sites throughout the study
area. The following shapefiles were added to ArcMap for navigation and sampling
purposes: the DEM of Osceola County, the DRG of Osceola County, roads, kettles, and
target kettle sites. The GPS unit (Garmin GPSmap 76C) was linked to an Argonaut
TFlex laptop for live position tracking, using software created by the Minnesota
Department of Natural Resources (DNR Garrnin software). For efficiency, target kettles
near road corridors were given preference for sampling. However, because road
conditions through the study area are poor (deeply incised two-tracks are common), some
of the key, target kettles that are far from roads needed to be navigated to on foot, using

the laminated topographic maps.

41

3.2.2 Addressing Hypothesis 1- Slopewash Origin

If the source of kettle bottom silt is the surrounding backslopes, brought in by
slopewash, then there should be a gradual increase in silt content from the kettle
backslope soils (the presumed source) to the kettle bottom deposits (site of deposition)
due to downslope particle size sorting (Kleiss, 1970; Malo et al., 1974; Chen et al., 2002).
A number of downslope sorting mechanisms, such as rainsplash, slopewash, and
sediment rafting (Paton et al., 1995), tend to preferentially transport finer sediments
further, and result in a predictable pattern of progressively finer soils downslope (Milne,
1936)

To determine whether or not slopewash was/is responsible for the silty deposits in
kettle bottoms, two field tasks were completed: 1) I measured the aerial extent of silty
deposits in kettle bottoms and 2) I collected two samples from each kettle- one from the
kettle bottom deposit and the other from about halfway up the steepest, closest concave
backslope position. The first task was completed to better understand the nature of the
boundary between the kettle bottom soils and the backslope soils; a gradual, diffuse
boundary would support a slopewash origin for the silt, whereas an abrupt boundary
would suggest that slopewash is unlikely. The second task was completed to better
understand the textural difference between kettle bottom deposits and kettle backslopes
across the study area. If the kettle backslopes contain silt that is of comparative amount
and texture to that deposited in kettle bottoms, then a slopewash origin is feasible. If the
silt content between the two landscape positions is drastically different, then it is likely

that slopewash is not responsible for the kettle bottom deposits.

42

3.2.2a Extent of silt in kettle bottoms

The aerial extent of kettle bottom silt was measured at two sites; these sites were
chosen because they represent a range of kettle shapes within the study area. One of
them was the largest kettle sampled (site 31), and the other was slightly smaller than
average size (site 17). Starting from the bottom center portion of the kettle bottom, the
top 20-25 cm of the soil was removed with a push probe, and hand textured. Hand
texturing continued at l m intervals along the kettle bottom, working outward (upslope).
When the sediment texture at this depth changed from silty to sandy, a stake was placed
in the ground. Then, the hand-texturing procedure started again from the bottom-center
of the depression. Eventually, an oval of 10-12 stakes outlined the boundary between the
silty kettle bottom deposits and the surrounding sandy backslope soils. This boundary
was easy to identify while push probing because the sandy soils in the backslopes tended
to fall out of the probe upon pulling it out of the ground, due to their low moisture
holding capacity. In contrast, the moister silty kettle bottom deposits tended to remain in
the push probe when pulled out of the ground. As such, the stakes were placed where the
soils became sandy and dry enough to easily fall out of the push probe. Finally, the long
and short axes of the ovals were measured, and their area was computed using the
formula for a simple ellipse. The ratio of the oval area (extent of silty deposits) to the
kettle area (automatically calculated during the kettle mapping effort) was then

determined for the two kettles.

43

3.2.2b Comparing texture of kettle bottom and backslope soils across the study area
Two samples were acquired from each kettle, using a bucket auger- one from the
kettle bottom and the other from the kettle backslope (Fig 3.1). The kettle bottom sample
was acquired from the bottom-center portion of the depression, at a depth where the
sediments were (as determined by hand texturing) the siltiest. The sample depth differed
in each kettle bottom, but ranged between 40-70 cm. The kettle backslope sample was
gathered from the closest, steepest, convergent backslope position (often about halfway
up the slope), given that this is where slopewash processes would be most concentrated
(Pennock and De Jong, 1987), and from which, potential sources of backslope silt would
be winnowed via slopewash. Kettle backslopes were consistently sampled from 15-30
cm depth in order to capture a representative sample that has not been stripped of its silt
content (if backslopes are indeed the source of kettle bottom silt). Of the 159 target sites
identified, 53 kettles were sampled, resulting in 106 kettle bottom and backslope samples

combined (Fig. 3.2).

 

 

Figure 3.1- Schematic representation of a typical kettle in the study area (not to scale) showing the
locations of samples gathered within each kettle.

.V I
059’”) 'L. .

, ' .u ..
p .1“ is - a, ..
. . .

 

Figure 3.2- Location of sampled kettles on the Evart Upland. Kettles are outlined red. A total of 53
kettles were sampled. Two samples were gathered at each site, resulting in a total of 106 samples.

3.2.3 Addressing Hypothesis 2- Aeolian Origin

The textural data from each kettle bottom deposit and adjacent backslope was
analyzed to address the feasibility of an aeolian origin for the kettle bottom silts at each
kettle site. If the textural data indicate that slopewash is not likely responsible for the
deposits of kettle bottom silt, then invoking an aeolian origin for the silt is reasonable.
The only other scenario that could explain the presence of silt in kettle bottoms is if the
kettles on the Evart Upland once contained ponded water, offering a low energy

environment in which near-surface silty sediment could be deposited. This scenario may

45

be feasible in some of the lower, wetter kettles close to the perimeter of the upland.
However, kettles on the majority of the upland would not likely pond water because they
are composed of coarse-textured, sandy outwash and are far above the current water
table.

If the silt in kettle bottoms does indeed have an aeolian origin, then the question
would remain as to why it was preserved there and not elsewhere on the landscape. Most
of the Evart Upland is topographically steep and sloping, offering few stable surfaces for
the preservation of aeolian silt (i.e., silt would be easily transported off of steep slopes
and eventually deposited into nearby kettle basins). However, there are some flat,
geomorphically stable landscape positions on the Evart Upland (albeit very few).
Geomorphically stable surfaces are defined herein as low-lying, flat landscape positions
that show no signs of recent erosion. These sites are more likely to collect and retain
aeolian silt than sloping surfaces because of the greater potential for runoff and erosion
inherent with increasing slope. Therefore, I chose three geomorphically stable surfaces to
sample to analyze their silt distribution with depth. Soils at these three sites were depth
sampled with a bucket auger at 10-20 cm intervals. Depth sampling continued down into
the C-horizon, ranging between 75-105 cm depth. An aeolian origin for the kettle
bottom silt would be supported by evidence indicating that geomorphically stable, flat
landscape positions contain an increase in silt in their upper profiles, since silt is more
likely to be preserved here than on sloping surfaces, and because aeolian silt would have
been the last sediments deposited, and thus would be retained in the upper profile. This

scenario assumes that a silt increase in the upper soil profile is not due to physical

46

weathering, which is reasonable for the Evart Upland, given that soils there are young

and relatively poorly developed.

3.2.4 Addressing Additional Research Questions

After the aforementioned field data were collected and analyzed, a number of
other research questions arose. These questions are outlined below with an explanation
of how additional field data were gathered to answer them. In all cases, the additional

data contributed supporting evidence for the origin of the kettle bottom silt.

3.2.4a - How do kettle bottom sediments change (texturally) with depth?

While sampling kettles across the study area, I encountered a number of kettle
bottoms containing buried soils. The presence of buried soils in kettle bottom deposits
revealed that this landscape has experienced alternating periods of stability and instability
through time. Clearly, the sequence of sedimentation in kettles would reveal more about
the nature of silt deposition in this landscape. Therefore, I chose eight kettle sites, each
containing buried soils, to sample with depth. I chose sites with buried soils because they
more clearly show signs of former landscape stability and instability than kettles without
buried soils. Using a bucket auger, samples were gathered at ~10-20 cm intervals. The
depth and thickness of buried soils, along with pedogenic horizon sequences, were noted
in the field. Depth sampling continued into the lowermost coarse sandy outwash
(substrate) sediments underlying the silt. The depth to these substrate sediments ranged

between 110 and 220 cm, depending on the site.

47

3.2.4b - When was silt deposited in kettle bottoms in the study area?

The opportunity to answer this question arose upon discovery of charcoal
fragments lodged in buried soils at two kettle bottom sites. In one kettle bottom
containing a particularly strong sequence of buried soils (site 33), a soil pit was
excavated. After cleaning the profile face, images of the soil were taken (see Fig. 4.38,
Chapter 4). The horizon sequence, depth and thickness of buried soils, and presence of
charcoal, were all noted. The lowermost buried soil contained a number of small
charcoal fragments (Fig 4.38, Chapter 4). A large volume (~25 liters) of this soil-with-
charcoal mixture was gathered and transported back to the soils lab at the Michigan State
University Geography building, to extract charcoal for radiocarbon dating (extraction
methods explained in next section). At another site (site 17), macroscopic pieces of
charcoal (about 0.5 to 3 cm in length) were discovered while depth sampling. These
charcoal pieces were carefully removed from the bucket auger shavings and placed in a

sample bag. The depth that the charcoal was recovered from was noted.

3.2.4c - What is the source of silt in kettle bottoms?

If the silt in kettle bottoms is aeolian, then it follows that perhaps it could be fi'om
some extra-regional source, and therefore, possess recognizable differences in silt
mineralogy, as compared to the silt within adjacent backslopes (local silt). On the other
hand, a similar mineralogical composition of silt in kettle bottoms and adjacent
backslopes would suggest a more local source for the kettle bottom silts. To answer this
question, the silt mineralogy of four kettle bottom and adjacent backslope pairs were

analyzed and compared using X-ray diffraction. Because the methods used to answer this

48

question are primarily lab-based, they are described in more detail below, in the section

titled “X-ray diffraction analysis.”

3.3 Laboratogy Methods

3.3.1 Sample Preparation

The soil samples were brought back to the soils lab in the Geography building and
allowed to air dry for several days. Samples that still contained moisture after several
days were heated in an oven at about 65 °C overnight to remove excess water. Each
sample was lightly ground with a mortar and pestle, and passed through a 2-mm sieve to
remove coarse fragments. The remaining fine-earth samples were then homogenized in a
sample splitter to ensure that a representative sub-sample for particle size analysis was

obtained.

3.3.2 Particle Size Analysis
3.3.2:! Background

In recent years, laser diffractometry has generally replaced the more traditional
sieve-pipette method for particle size analysis (Sperazza et al., 2004; Arriaga et al.,
2006). Laser diffraction analysis is faster, more accurate, and provides a near-continuous
distribution of particle size data, whereas data generated by the sieve-pipette method
generally are output as discrete particle size fractions, e. g., 50-125 microns, 2-50
microns, etc. Although the data produced by the two methods are comparable and highly
correlated (Arriaga et al., 2006), some differences do exist, mainly in the clay fi'action

(Buurman et al., 1997).

49

Laser diffractometry commonly underestimates the amount of clay (< 2 microns),
when compared to the pipette method (Loizeau et al., 1994; Beuselinck et al., 1998)
probably because of its platy shape, which slows the rate at which it settles within a
traditional pipette analysis cylinder. For this reason, Konert and Vandenberghe (1997)
suggested that a clay-silt break of 8 microns be utilized for laser diffractometry data, in
order to facilitate any comparisons between laser diffractometry and traditional pipette
particle size analysis data. Similarly, in-house data from the soils lab at Michigan State
University suggested that correlations between clay contents by pipette (from former
studies) vs. laser diffractometry are highest when the clay-silt break for the latter is set at
6 microns. Thus, in this study, the clay/fine silt boundary was set to 6 um to offset the

slight over-estimation of clay by the laser hardware.

3.3.2b Data output

Particle size distributions for each sample were measured with a Malvem
Mastersizer 2000 laser particle size analyzer. This particular model measures the
distribution of grain sizes from 0-1000 pm in near continuous detail, while the 1000-2000
um fraction is expressed as a single value. The user defines what data are included in the
output file, depending on the intent of the research. For this study, two output templates
were used: the first includes the necessary fields to create a particle size distribution
curve, the second includes many grain size ranges, including clayfree calculations and

ratios of certain silt fractions. These data tables are listed in the Appendix.

50

3.3.2c Obtaining a representative sub-sample

Before the particle size analysis was run, a representative sub sample from each
sample bag was placed into a small vial. The quantity placed in each vial varied
depending on the texture of the sample (very sandy samples require more sediment,
where clay or silt rich samples require less sediment), but, in general, the weights ranged
between 0.25 to 1.0 g. A mixture of distilled water and 5 mL of dispersing solution
((NaPO3)6 and NazCO3) was added to each vile. The vials were then agitated on an

orbital shaker for two hours to fully disperse the individual grains in the sample.

3.3.3 Radiocarbon Analysis
3.3.3a Sample Preparation

The charcoal removed from kettle bottoms at site 17 and site 33 required different
preparation techniques for radiocarbon analysis because the fi'agments were much
different in size. The larger charcoal pieces gathered from site 17 required very little
further preparation; they were removed from the sample bag with tweezers, rinsed with
distilled water, and placed in a beaker to dry in an oven at ~25 °C overnight. The
smaller charcoal pieces within the bulk-sampled buried soil from site 33 were extracted
as follows: aliquots of soil were rinsed through a 250 um sieve to separate the mineral
component from the larger fragments of charcoal. The charcoal fragments (and other
sediment larger than 250 pm, which included plant fragments and sand grains) were then
rinsed with distilled water and carefully flushed onto a petri dish. The petri dish was then
placed under a Leica MZ6 binocular microscope and pieces of the charcoal were

removed with tweezers and placed in distilled water and dried in an oven at 65 °C until

51

all the water evaporated. This process- sieving, extracting, and drying- was repeated until
the desired dry weight of 0.2 grams total charcoal was obtained. The samples were
repackaged in new sample bags, labeled, and sent to the Center for Applied Isotope
Studies (CAIS) at the University of Georgia. There, the Graphite l4C/13C ratios of the
charcoal samples were analyzed using an accelerator mass spectrometer. The '3 C/ 12C
ratios were measured separately using a stable isotope ratio mass spectrometer and

expressed as 513C with an error of less than 0.1%o.

3.3.4 X-Ray Diffraction Analysis
3.3.4a Silt separation

The goal of the silt separation procedure that I employed in this research was to
isolate a range of grain sizes shared by adjacent kettle bottom/backslope sample site
pairs. The four sites chosen, from among 53 total pairs, included those that showed the
greatest evidence for an aeolian origin for the kettle bottom silt samples*- sites 24, 28, 31,
and 42. At sites 28 and 31, the kettle bottom silt appeared least likely to be derived via
slopewash (i.e. these sites had the greatest difference in silt content between the kettle
bottom and adjacent backslope). Sites 24 and 42 were kettles included in the depth
sampling procedure. These two sites contained intervals of well-sorted, nearly pure silt
in the sequence of kettle bottom sediments. It is assumed that these intervals most clearly
show evidence of an aeolian depositional unit in the kettle bottoms, and therefore,
mineralogical differences (if evident) between them and adjacent backslope silts would

be most pronounced in such samples.

 

' This procedure was carried out after examining the results of particle size analysis.

52

The 20-53 pm gain size range was chosen for the silt separation analysis. The
lower limit (20 um) was chosen because the four backslope samples show an increase in
silt at 20pm, and contain very few particles below that size (Fig. 3.3). The upper limit
chosen for silt separation was 53 um, given that this size is the closest sieve size available
to the upper range of silt (50 um). The 20-53 pm size fraction was separated fi'om kettle
bottom and backslope samples as follows: approximately 10 garns of each sample were
placed in a bottle containing a mixture of distilled water and 5 mL of dispersing solution
((NaPO3)6 and NazCO3). The bottles were agitated on an orbital shaker for two hours to
fully disperse the sediments. The dispersed samples were passed through a 53 um sieve
with distilled water, into a 1000 mL gaduated cylinder, to separate any sediments larger
than 53 um. Sediments smaller than 20 um were isolated using a calculation of Stokes’

law that determines the time it takes for gains to settle to a given depth. Using this law, I

 

   
  
  
  
 

 

 

 

 

 

 

9 1 f
. Clay Silt Sand
8
— Backslope 28
7 _ —— Backslope 31
6 ‘: — Backslope 24
\° :
e '1
p 5 1 Backslope 42
= -l
'5 4 i
> _
3 -I
2 €
1 -Z
3
O _' I I r l l l l l I
0.5 1 10 100 1000

Particle size (microns)

Figure 3.3- Particle size distribution curves for selection of backslope samples used in the X-ray
diffraction analysis. The range of grain sizes extracted from the samples is 20-53 pm. The lower
limit of 20 um was chosen because backslope samples have very little silt below that size.

53

determined that sediments larger than 20 um should settle at or below 20 cm alter three
minutes, 48 seconds. To facilitate this separation procedure, I then mixed the samples in
the cylinder by tipping them back and forth multiple times to fully suspend the sediments.
Then, the sediments were allowed to settle in the gaduated cylinder for 3 minutes, 48
seconds, after which time a suction hose was used to remove the sediment-rich
supernatant to a depth of 20 cm. This process was repeated 4-7 times, depending on the
clay content of the samples, until the supernatant was clear.

The 20-53 pm sub-samples were then dried in an oven overnight at 65 °C and
then pulverized with a Fritsch Analysette 3 Spartan Pulverisette for about three minutes
until the silt was uniformly crushed to < 5 pm. The crushed silt powder was then stored
in vials for transport to the X—Ray diffraction lab in the department of Geological and

Environmental Sciences at Hope College, in Holland, Michigan.

3.3.4b Lab analysis

A total of eight samples were analyzed (four kettle bottom/backslope pairs). A
random powder mount technique was used for the X-ray diffraction analysis. This
technique involved carefully tapping the powder into the container, using the edge of a
glass slide in order to optimize random orientation of the crushed silt particles (Zhang et
al., 2003). The randomly oriented silt powder was then X-rayed in a MiniFlex+ X-ray
Diffractometer (Rigaku Corporation, The Woodlands, TX) using Cu-Ka radiation. The
first three samples were X-rayed in the 21°-33° 20 range for a two second count time and
a 002° step size, as an exploratory method to determine what minerals were present. It

was later concluded that the primary peaks for any siglificant minerals present in the

54

samples were captured within the 25°-29° 20 range, and therefore, the remaining samples
were X-rayed within this range. Quartz, K-feldspar, and plagioclase were the only three
clearly identifiable, commonly occurring minerals present in the samples. The three
minerals were assigned to 26.5 20, 27.4 20, and 27.8 20 values, respectively. The
diffraction patterns of the four kettle bottom-backslope pairs were plotted, stacked, and

qualitatively compared to look for significant differences in silt mineralogy.

55

CHAPTER 4 - Results and Discussion

Fieldwork and sampling confirmed that 53 of the 59 kettles visited in the field
contain lenticularly-shaped, silty deposits in the center-bottom portion (i.e., the deepest
part) of the depression. The recurring pattern of silt deposition in kettle bottoms
throughout the Evart Upland indicates that kettles here have acted as accumulation
basins, preserving a record of landscape stability and instability since deglaciation, via
the sediments that are contained within them, particularly at their center-bottom areas
(Walker and Ruhe, 1968). The purpose of this chapter is to explain the characteristics
and likely origin(s) of these silty kettle bottom deposits, by describing their textural
variability at various sites across the upland, as well as with depth at specific sites. The
paleoenvironmental significance of such deposits, in the context of the regional

landscape, will be discussed in Chapter 5.

Addressing Hypotheses- Slopewash vs. Aeolian Origin for the Kettle Bottom Silt
If slopewash were solely responsible for the accumulation of silt in kettle
bottoms, then two conditions could be assumed to exist:
1. Silty kettle bottom deposits should have a diffuse lateral boundary, gadually
transitioning up to the surrounding sandy kettle backslopes, and
2. Backslopes should contain siglificant amounts of silt that is of comparable
particle size distribution (i.e., similar distribution of particle sizes within the silt
[2-50 micron] fraction) to that in kettle bottoms, bearing in mind that the subtle

sorting effect of downslope transport may skew the silt distributions slightly.

56

If these conditions do not hold, than it is reasonable to invoke an aeolian origin
for the kettle bottom silt, because there are no other reasonable mechanisms by which
such silty deposits could consistently occur in kettle bottoms in this landscape. The only
other mechanisms to consider that could explain the origin of the silty sediments in kettle
bottoms involve increased weathering of in situ sediments there, or deposition of silts in a
subaqueous (i.e., a kettle lake), setting. However, both of these scenarios are unlikely
because 1) the soils here are relatively young, and as such, have not weathered deeply,
and 2) the soils in the sampled kettles are well drained, with deep water tables, and as
such, have not likely ponded water in the past.

In this section, I will first discuss the nature of the outer boundary (edge) of kettle
bottom deposits at two representative sites. Next, I will explain the textural
characteristics of the kettle bottom deposits and their respective kettle backslope
sediments. Finally, I will examine the textural difference between kettle
bottom/backslope pairs at nine individual kettle sites, highlighting examples where the
silty kettle bottom deposits exhibit the most, average, and least textural differences from

their presumed upslope source (i.e. backslopes).

4.1 Extent of Silt in Kettle Bottoms

The outer boundary (edge) of the silty kettle bottom deposits was examined in
detail at two kettles (site 17 and site 31). These sites represent the largest (site 31) and
average (site 17) size kettles on the upland. I assumed that if the silt in kettle bottoms
was winnowed out of the surrounding backslopes via slopewash, then the outer boundary

of the silty kettle bottom deposits should be diffuse, extending up into the surrounding

57

backslopes. However, at both sites, the outer boundary of the kettle bottom deposits is
marked by an abrupt transition (~l-2 m lateral distance along the gound) from silt-
dominated sediments to sand-dominated sediments. The outer boundary was outlined in
the field by placing stakes in the gound where the surface sediments became dry, sandy,
and easily fell out of the push probe (see Methods). This outline of stakes crudely
delineated the extent of the silty deposits at both sites, and confirmed that the silty
deposits in both kettle bottoms are confined to (and concentrated in) the bottom-center
portion of the depression, and have an abrupt lateral boundary (Figs 4.1 and 4.2).

In summary, by examining the outer boundary of the silty kettle bottom deposits
at two representative sites, the following characteristics were confirmed:

1) The silty kettle bottom deposits are confined to, and concentrated in, the

center-bottom portion of the basins.

2) The boundary between the silty kettle bottom deposits and the sandy

backslope sediments is reasonably abrupt.

3) The silty deposits do not extend up the surrounding backslopes.

These combined characteristics suggest that the silty kettle bottom deposits lie
unconformably within kettle bottoms, and as such, appear anomalous with respect to
gadual downslope sorting of backslope sediments (i.e. slopewash). Furthermore, the
silty deposits comprise very little of the area of the kettle bottom (Figs. 4.1 and 4.2).
Therefore, condition 1 (listed above) does not hold true for the two representative sites
examined in the field. The texture of backslopes and kettle bottom deposits will be

explored next to examine condition 2.

58

  

 

 
 
 
 

Kettle site 17

   

Kettle area = 21,300 sq. m ,.

Area of silt in kettle bottom = 34.6 sq. m

% of kettle occupied by silt = 0.17%
. , 9 . .

  

 

 

 

 

 

 

 

Area of silt in kettle bottom = ABrr

A=4.04m
B=2.73m

Area = (4.04)(2.73)1'r
Area = 34.6 sq. m

 

Figure 4.1- Aerial extent of silty sediments in
the bottom of kettle site 17 (above). The kettle
area is shaded gray. The extent of silt is shaded
brown. The red arrow gives the perspective of
the photograph at left.

 

 

 

 

59

 

       
   

Ke__tt_l___e site 31

Kettle area= 63 880 sq m
Area of silt In kettle bottom: 90 8 sq. m
% of kettle occupied by silt- - 0.14%

11...}

 

 

 

 

 

 

 

 

Area of silt in kettle bottom = ABn

A=6.25 m
B =4.63 m

Area = (6.25)(4.63)n
Area = 90.8 sq. m

 

Figure 4.2- Aerial extent of silty
sediments in kettle bottom at kettle
site 31 (above). The area of the
kettle is shaded gray. The extent of
silt is shaded brown. The red
arrow gives the perspective of the
photograph at left.

 

 

 

 

60

W

In support of the evidence presented above regarding the nature of the outer
boundary of silty, kettle bottom deposits, samples were gathered from kettle bottoms and
adjacent backslopes (kettle bottom/backslope pairs) at 53 kettle sites across the study
area. The texture of kettle bottom deposits and surrounding backslope samples were then
analyzed. Their textural components were compared, at a site-by-site basis, to determine
if the backslopes contain siglificant amounts of silt of comparable particle size
distribution to that deposited in kettle bottoms. First, I will explain the general, average
textural components of kettle bottom deposit and backslope samples (herein referred to as
“sample types”). Then, I will explain the particle size distribution curves of each sample
type at specific sites to compare their modal sand and silt components, and use these data

to examine the feasibility of a slopewash origin for the silt in kettle bottoms.

4.2.1 General Texture

Results of the textural analyses revealed a number of important textural
characteristics of both sample types (Table 4.1). The most important textural
characteristic, with respect to the feasibility of a slopewash origin for the kettle bottom
deposits is that, on average, kettle bottoms contain ~four times the amount of clayfree silt
than do backslopes. The drastic textural contrast between the two sample types suggests
that 1) kettle bottoms are geatly enriched with silt relative to backslopes, and thus, 2)

backslopes do not contain enough silt to be the primary source of kettle bottom silt.

61

Table 4.1- Average textural components of backslope and kettle bottom deposit samples.

 

 

. %Clayfree % Clayfree
*
Sample Type % Sand (0) % Sllt (o) % Clay (0) sand (6) silt (o)
Backslopes 82.6 (7.8) 14.3 (6.1) 3.1 (2.0) 86.2 (6.7) 14.8 (6.7)
Kettle
bottoms 36.1 (14.5) 48.6 (10.5) 15.3 (5.1) 41.9 (14.5) 58.1 (14.5)

 

 

 

 

 

 

* 0 denotes standard deviation.

On average, both sample types are texturally much different from each other
(Table 4.1). A paired t-test revealed siglificant differences (p < 0.001) between each
sample type for all textural components listed in Table 4.1. Individual samples within
each sample type exhibit sigrificant textural variability. Figs. 4.3 and 4.4 show the
particle size distribution curves of all 53 kettle bottom and backslope samples,
respectively. These figures illustrate that kettle bottom deposit samples have geater
textural variability than do backslope samples. Many kettle bottom deposits contain
significant amounts of sand", and some backslopes contain sigrificant amounts of silt,
thereby deviating from their average “silty” or “sandy” textures, as reported in Table 4.1.
In effect, there are a number of backslope samples that texturally “overlap” with kettle
bottom deposit samples (Fig. 4.5). Hence, some backslope sites may be a feasible source
for silt accumulation in their respective kettle bottoms (via slopewash), other things being
equal. However, the vast majority of other kettle sites exhibit a drastic textural contrast
between the two landscape positions; the backslopes do not contain a comparable amount
of silt to that deposited in kettle bottoms. Thus, at these sites, the backslopes could not be

the source of kettle bottom silt.

 

' The potential sources of the sand in kettle bottom deposits will be discussed later in the chapter.

62

 

—n
N

Clay Silt Sand

p—s
p—s

Volume%
—0
Us 0‘ \l 00 \O O
llLllLiLllllllllllllliJLLLLJlllllJJlLlLJIllllll

 

I—‘NW-h

 

 

llllllJ_LI_Llll

O

 

I
r11 ITITI

100 1000

0
UI
~
~
O

Particle size (microns)

Figure 4.3- Particle size distribution curves of all 53 backslope samples. The average particle size
distribution is highlighted red. Blue and green curves represent the coarsest and finest textured
sample, respectively.

63

 

y—d
O

\l
lllrlrrerlmLarhrthrmllllrlLJLL

Clay Silt Sand

\0

00

Volume %
A Ur 0\

DJ

 

N

lllLllllllllllllll

fl

O

 

 

 

.9

Particle size (microns)

Figure 4.4- Particle size distribution curves of all 53 kettle bottom deposit samples. The average
particle size distribution is highlighted red. Blue and green curves represent the coarsest and finest
textured sample, respectively.

64

 

 

   
  

   

Clay content 80% Silt content
20%

 

100/ * silt loam 90%
0

l 00%
0%

 

\ \ V
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

4—— Sand content

Figure 4.5- Lower half of the USDA textural triangle, showing the texture of kettle bottom samples
(blue star) vs. kettle backslope samples (green x). Kettle bottom deposit samples generally have loam
and silt loam textures. Kettle backslope samples are dominantly sand and loamy sand in texture.
The main area of textural overlap of kettle bottoms and backslopes is circled and shaded gray.

65

To better clarify if slopewash is responsible for transporting silt to the kettle
bottoms, the modal silt components of representative kettle bottom/backslope pairs was

also compared, using more detailed particle size distribution data.

4.2.2 Particle Size Distribution Curves

Particle size distribution (PSD) curves (Figs. 4.3 and 4.4) show the distribution of
gain sizes in a particular sample; particle size is represented on the x-axis, whereas the
volume % of the sample within a given particle size range is represented on the y-axis. In
actuality, the data for particle size distribution are composed of closely spaced (small x-
axis binning value) x-y pairs, and the spacing of the binning value increases with
increasing particle size. As a result, the particle size distribution data have a higher
resolution in the smaller gain size ranges (i.e. clay and silt). Because the binning values
are so closely spaced, the data are commonly plotted using a connecting line so that the
result resembles a continuous curve (Beierle et al., 2002), which is somewhat misleading
vis a vis the actual data. Traditionally, particle size distribution curves are plotted
logarithmically, to highlight the details of smaller gain size ranges. The location of a
curve peak, with respect to the x-axis, represents the most frequently occurring gain size
(modal gain size) of the sample. If a sample contains sigiificant amounts of multiple
gain size components (such as sand and silt), it may contain a peak for each, and
therefore, is considered bimodal. Many sample PSD curves in this research are bimodal,
and as such, peaks will be herein referred to as “silt peaks” and “sand peaks” when and if

they fall within the silt and sand fractions, respectively.

66

4.3 Comparing Silt Content in Kettle Bottom and Backslope Pairs

4.3.1 Average PSD Curves of All Kettle Bottoms and Backslope Samples

Fig. 4.6 shows the average particle size distribution (PSD) curves for kettle
bottom and backslope deposits. The two curves are distinctly bimodal. The most
prominent peak for kettle bottom deposits is the silt peak, centered at 18 um, whereas the
minor peak is in the sand range at 315 pm. In contrast, the prominent peak in kettle
backslope samples is in the sand range at 315 um, and the minor peak is in the silt range
at 45 pm. These data show that the location of the sand peak (along the x-axis) is similar
for both sampled populations, whereas the silt peak is drastically finer for kettle bottom
deposits. The shared sand peak location in both curves suggests that the sandy
component of kettle bottoms and backslopes is from the same population. In contrast,
however, the difference in silt peak location for both curves suggests that the silty
component of kettle bottoms and backslopes is from different populations.

The data in Fig. 4.6 suggest that slopewash may be responsible for contributing
some sand to kettle bottoms, but it is not responsible for contributing finer modal silt (i.e.
finer silt may have been transported into the kettle bottom from outside the basin). If this
interpretation is correct, then backslopes may have contributed some silt to kettle bottoms
via slopewash (albeit limited in quantity and coarser overall), but are not likely the source
of the finer modal silt that dominates kettle bottom deposits. This interpretation will now
be applied to data from a selection of kettle sites that contain the most, average, and least

texturally different kettle bottom-backslope pairs.

67

 

—l
O
l

: clay silt sand
9 1 Z I

Z 315
8 _ pm
7

Volume%
.h VI 0\
IllllLlllmllllllllllllllLJlllllll

L»

N

 

fl

 

 

 

lrllll.

O

 

IIIIIII] I T— T.‘IIIIII I I—FIIIITI

10 100 1000
Particle size (microns)

.9
(It
p—a

 

— Backslopes

— Bottoms

 

 

 

Figure 4.6- Average particle size distribution curves for all kettle bottom (red) and backslope (blue)
samples. Both populations have a bimodal particle size distribution with peaks in the sand and silt
range. The silt peak is different for each population (18 pm and 45 pm for bottoms and backslopes,
respectively), while the sand peak is the same (315 pm).

68

4.3.2 Most Texturally Different Kettle Bottom/Backslope Pairs

Figure 4.7 Shows three of the most texturally different kettle bottom/backslope
pairs (i.e., the geatest difference in silt contents between both sample types). The three
kettle bottom deposits average 51 % more silt than their respective backslope samples.
Similar to the average PSD curves in Fig 4.6, the backslope samples have a minor peak in
the silt range, but this peak is much coarser and less prominent than the corresponding silt
peak in kettle bottom samples. The difference between the two silt peak locations (with
respect to the x-axis) for adjacent sample types will be referred to herein as “silt peak
spacing”. The data in Fig. 4.7 support the notion that 1) kettle bottom deposits are
enriched with silt relative to the backslopes, and 2) backslope deposits do not contain
enough silt to be the primary source of kettle bottom silt. Perhaps most importantly, the
data in Fig. 4.7 illustrate that the silt in these three kettle bottoms is much finer than the

silt in backslopes, and therefore, probably not derived from the backslopes at all.

4.3.3 Average Textural Difference Between Kettle Bottom/Backslope Pairs

Figure 4.8 shows data from three kettle bottom/backslope pairs that represent the
average difference in silt content between the two landscape positions. The PSD curves
of backslope samples in this gaph are very similar to those in Figure 4.7. The spacing of
silt peaks in one of the kettle bottom/backslope pairs is slightly larger than in the previous
example (23 rim->26 pm), whereas the other two are the same (35 um and 45 pm).
Therefore, the primary difference between these samples and the previous samples is that
the kettle bottom deposit samples in this gaph contain more sand. The data in Fig. 4.8

support the interpretation that slopewash processes may be responsible for

69

Silt peak spacing (microns) between bottoms/backslopes
0.5 1 10 100 1000

 

 

 

 

 

 

 

 

9 lIIrI r i 1 114111 1 1 i-llLlll r r ianmr
I — Bottoms 28 |-——! 23 um ff.
81: ------- Backslope 28 I—-—; 35 um 5:353,
I ' i '-.
7 1 — Bottom 31 l :45 pm if
3 ------- Backslope 31 :5
6-3 — Bottom 34
°\o 5 j ------- Backslope 34
o .
a -
3 I
a .—
> 4 3
3%
2‘.
l-i
_ ”unu'flunutliiili
O_ I'] A‘ I W I I I I I I I T I I I I T L IT I I I I—r I
0.5 1 10 100 1000

Particle size (microns)

Figure 4.7- Particle size distribution curves for three kettle bottom/backslope pairs that have the
greatest textural difference (i.e. have the greatest difference in silt content) of the population of
sample pairs. The kettle bottom/backslope pairs are color-coded; solid lines represent kettle bottom
samples, while the dashed lines represent corresponding backs10pe samples. The colored bars at the
top of the graph indicate the difference in silt peak location between bottom/backslope pairs.

70

Silt peak spacing (microns) between bottoms/backslopes

 

 
  
  
  
 
   
 

 

 

 

 

 

0.5 l 10 100 1000
9‘....l . .l.....l . .lllllll A la...”
. — Bottom33
8-: - ----- - Backslope 33
j — Bottom22
7':
a ------- Backslope22
6; — Bottom29
°\o 5: ------- Backslope29
o -
S 3
o 4
>4:
3-1
2-3
I-i
0.‘
05 10 100 1000

Particle size (microns)

Figure 4.8- Particle size distribution curves for three kettle bottom/backslope pairs representative of
the average difference in silt content among the whole population. The kettle bottom/backslope pairs
are color-coded; solid lines represent kettle bottom samples, while the dashed lines represent
corresponding backslope samples. The colored bars at the top of the graph indicate the difference in
silt peak location between bottom/backslope pairs.

71

contributing some sand to kettle bottoms, but that the majority of the Silty component is

much finer, and therefore, probably not winnowed out of the backslopes via slopewash.

4.3.4 Most Texturally Similar Kettle Bottom/Backslope Pairs

Figure 4.9 shows data from the three kettle bottom/backslope pairs that exhibit the
geatest similarity in texture. These samples have the least difference in silt content
between the kettle bottom and adjacent backslopes. The kettle bottom samples in this
gaph contain more sand than any other kettle bottom samples. Interestingly, this subset
of samples contains both the largest (kettle site 26) and smallest (kettle site 3) silt peak
spacings between corresponding kettle bottom/backslope pairs, as compared to any of the
previously shown samples (Fig. 4.9). The small silt peak spacing (and overall similarity
of the PSD curves) for kettle site 3 may indicate that the two samples are from the same
population (Fig. 4.9). Perhaps, the kettle bottom sample at site 3 was gathered from a
depth interval dominated by an influx of sand from the surrounding slopes (i.e., a
slopewash interval) *. The data in Fig 4.9 again suggest that slopewash is probably
responsible for transporting some of the sandy sediment to the kettle bottom from the
surrounding backslopes, but that slopewash is not likely responsible for the enrichment of
finer modal silt in kettle bottom deposits. Such a conclusion is best expressed in the PSD
curve data for site 26 (Fig. 4.9), where the large silt peak spacing indicates that, despite
the preponderance of sand (possibly from the surrounding backslopes), silt of a different
origin is nevertheless detectable in the kettle bottom deposit because the mode gain size

of the silt peak is much finer.

 

‘ The depositional sequence of eight kettle bottom deposits will be explored later in the chapter to better
understand their textural variation with depth.

72

Silt peak spacing (microns) between bottoms/backslopes
1 10 100 1000

104 ill 1 l rlrrrl r IILLILIJ l 1 111411
- — Bottom3

 

  
  
  
  
  
   
  

9 T ------- Backslope 3

8 ‘j — Bottom 4

3 ------- Backslope 4
1 — Bottom 26

6 _ ------- Backsope 26

 

 

Volume %
A U!

lllllllLL¥lLlLlJlllLlilel

b.)

 

N

y—a

O

 

 

 
  

IIIIII

1 0
Particle size (microns)

.9
u:
fl

Figure 4.9- Particle size distribution curves for three kettle bottom/backslope pairs that are the most
texturally similar. The kettle bottom/backslope pairs are color-coded; solid lines represent kettle
bottom samples, while the dashed lines represent corresponding backslope samples. The colored
bars at the top of the graph indicate the difference in silt peak location between bottom/backslope
pair.

73

4.3.5 Summary of Textural Data

4.3.5a Silt contents

As shown above (Figs. 4.6-4.9), kettle bottom deposits contain more silt and finer
modal silt contents than their corresponding backslope pairs. If slopewash alone were
responsible for the silt in kettle bottoms, then the silt portion of the PSD curves for
adjacent kettle bottom and backslope samples would have nearly similar silt peak
locations, or perhaps slightly finer silt peaks in kettle bottom samples due to downslope
sorting. The difference in silt peak locations suggests that the silt in kettle bottoms was
not likely winnowed out of the surrounding backslopes via slopewash. Furthermore, in
the context of slopewash processes, it is unlikely that fine silt would be transported in the
absence of coarse silt, as the PSD data suggest (Figs. 4.6-4.9).

As stated earlier, another reasonable mechanism by which fine silt could be
deposited in this landscape is if kettles here ponded water at some point in their
evolution, offering a low energy environment for the deposition of silt. However, this
scenario is not likely, given that the substrate sediments underlying the kettle bottom
deposits are well drained with deep water tables. As such, condition 2 (listed above) does
not hold true for these samples, and therefore, it is likely that the majority of silt in

kettle bottoms is aeolian.

74

4.2.5b Sand contents

Sigrificant variation with respect to sand content exists in the kettle bottom
deposits. Some of the kettle bottom deposits contain very little sand, whereas others have
bimodal PSD curves, containing a significant sand peak (Fig. 4.4). Because the modal
sand components of backslopes and adjacent kettle bottom samples are similar (Figs. 4.6-
4.9), it was previously suggested that slopewash might have been responsible for
contributing sand to the kettle bottoms (but not finer modal silt). As it has been
suggested that the majority of silt in kettle bottoms is aeolian, it must be acknowledged
that perhaps some of the sand in kettle bottoms is also the result of aeolian deposition,
rather than via slopewash. Furthermore, it is also possible that sand was brought up from
the sandy substrate sediments that underlie the silty kettle bottom deposits via
bioturbation. As such, sand in kettle bottom deposits may be attributed to three possible
sources: 1) slopewash sand from the surrounding backslopes, 2) aeolian sand that was
locally redistributed and deposited in the kettle bottoms, or 3) in-mixed sand from the
underlying sandy substrate sediments, brought upward via bioturbation. The scenario
favored herein is that the surrounding backslopes have contributed sand to kettle bottom
deposits via slopewash deposition. Such a scenario seems most likely because it is the
simplest explanation; nearly all of the sampled kettle bottoms have steep, sandy slopes
surrounding them on all sides, which would have been susceptible to downslope transport
during unstable periods. If so, then downslope sorting would likely result in a slightly
finer mode sand gain size (sand peak) in kettle bottom deposit samples, as compared to
backslope samples. To test this hypothesis, the modal sand components of adjacent

samples at various kettle sites were compared.

75

Table 4.2 lists the sand peak locations for adjacent samples (backslope-)kettle
bottom), at a site-by-site basis, for all kettle bottom deposits that contain a recogrizable
sand peak in their PSD curve. In Table 4.2, sites are labeled “C”, “F”, or “-”, to denote
“coarser downslope”, “finer downslope”, or “no change” with respect to the sand peak in
backslopes and adjacent kettle bottom deposits. To visually represent the data in Table
4.2, vectors showing the amount and direction of downslope changes in mode sand gain
size were plotted (Fig. 4.10).

Overall, the sand peaks (mode sand gain size) in kettle bottom deposits nearly
match the corresponding sand peaks in adjacent backslope samples, suggesting that the
sand in kettle bottoms was derived from the surrounding backslopes, or similar sediment
(Table 4.2). At 20 of the 32 sites examined, the sand peak in kettle bottoms is either the
same or slightly finer than that in backslopes, as would be expected due to downslope
sorting. In contrast, however, at 12 of the 32 sites, the modal sand gain size is coarser in
kettle bottom deposits than in the backslope samples. It is difficult to interpret the
“coarser downslope” data, other than they do not appear to fit with traditional notions of
downslope particle size sorting along hillslopes (Malo et al., 1974; Chen et al., 2002).
Furthermore, the data in Table 4.2 indicate that, at 8 of the12 “coarser downslope” sites,
the sand peak location is only one x-axis bin size coarser (on the PSD curve), and
therefore, may be insignificant. In summary, data from 20 of the 32 Sites illustrate that
slopewash is probably responsible for contributing some sand to kettle bottoms, but data

from the remaining 12 sites do not support such a conclusion.

76

Table 4.2- Downslope, point-to-point comparisons of mode sand grain size.

 

For all kettle bottoms with a prominent sand eak (bimodal distribution)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

S Modal sand grain Volume % of Modal sand grain Volume % Of Amount of
ample . . . . sample at
site srze in backslope sample at .mode srze rn kettle bottom mode (peak C, F, or - ? change
(um) (peak helght) deposxt (um) height) (um)
I 363 5.92 275 3.86 F -88
3 315 5.89 363 4.92 C 48
4 363 6.38 239 5.86 F -124
5 315 7.35 363 5.75 C 48
7 315 9.56 315 3.71 - -
8 315 8.10 413 3.15 C 103
9 315 7.94 315 2.00 - -
10 315 4.83 315 1.42 - -
11 315 9.93 315 4.54 - -
12 275 8.77 413 1.74 C 138
19 363 6.67 275 2.37 F -88
20 315 6.25 315 0.40 - -
22 363 8.89 239 3.51 F -124
23 315 8.95 208 2.10 F -107
24 315 7.90 158 2.11 F -157
25 275 7.99 275 5.03 - -
26 315 9.70 363 9.17 C 48
27 478 8.78 413 4.97 F -65
29 315 7.46 363 2.45 C 48
30 315 7.91 363 2.57 C 48
31 413 8.91 275 1.18 F -138
36 315 6.63 275 2.34 F -40
38 315 6.18 275 0.86 F -40
40 363 8.70 239 2.73 F -124
41 315 7.90 275 2.38 F -40
43 275 3.39 315 1.59 C 40
44 363 8.00 413 1.12 C 50
46 363 7.36 275 3.03 F -88
47 315 9.79 275 2.29 F -40
48 315 8.16 363 1.77 C 48
55 363 5.91 413 2.63 C 50
56 315 7.53 363 2.46 C 48
Average = 33 1 pm Average = 315nm 2:11: 31:21:15:
Mode = 315nm Mode = 276nm

(-) = 6 -17 pm

77

 

 

— Coarser mode sand grain size downslope — Finer mode sand gain size downslope

 

 

 

 

 

 

 

 

 

 

 

 

10 10
I \s I
84 3-
.l
.\‘ 6-a 6—
0 r a
E 4
:I
B -t
> 4? 4‘1
2. \ .
_ 2-
n=12 -
.
0 III IIII IIII 'IITII 0 I‘ll TIII TIIIII'II IIIIIIIII

 

. . I . . I . . . .
150 200 250 300 350 400 450 150 200 250 300 350 400 450
Mode Sand Grain Size (microns)

Figure 4.10- Vectors showing the amount of downslope coarsening (blue) and fining (red) with
respect to mode sand grain size between backslopes and kettle bottoms. Each vector contains two
sets of x-y coordinates. The starting position of the arrow is the sand peak in backslope samples.
The end position of each arrow represents the sand peak in the adjacent kettle bottom. Large
changes in the x-coordinate denote significant coarsening or fining downslope. Large changes in the
y-coordinate denote significant differences with respect to the amount of sand between the two
landscape positions.

78

A reconnaissance trip to the field revealed direct evidence for a recent slopewash
event at kettle site 33 (Fig. 4.11). At this site, a gully has been eroded down the
backslope, resulting in a fan-like deposit downslope in the kettle bottom. This erosional
event is preserved in the kettle bottom deposit as a layer of sandy sediment overlying the
uppermost buried soil (Fig. 4.12). The uppermost buried soil in Fig. 4.12 has abrupt
lower and upper boundaries, and contains abundant charcoal, suggesting that this
erosional event may have been triggered by, or directly followed, a fire. It has long been
recogrized that fire is responsible for causing temporary hydrophobicity in recently
burned soils, thereby facilitating runoff in otherwise well drained sediments (Savage,
1974; DeBano et al., 1979). In that context, backslope deposits are probably the source
of sand in kettle bottom deposits, especially where there has been direct disturbance due
to fire*.

The textural evidence presented thus far suggests that the kettle bottoms have
collected aeolian silt, as well as varying amounts of slopewash-derived sand, which have
been redistributed to the bottom-center portions of the kettle basins. Because these two
types of sediment are drastically different in the context of depositional processes, the
aeolian silt may have been deposited in the kettles at different times than slopewash sand,
leading to textural heterogeneity with depth in the kettle bottom deposits. Furthermore, it
is possible that bioturbation has, to varying degees, mixed slopewash-derived sand
throughout the silty kettle bottom deposit profiles. Therefore, the textural variation in

kettle bottom samples (Fig 4.4), particularly with respect to sand content, is probably

 

' The role of fire in destabilizing soils in the Evart Upland will be explored later in Chapter 5.

79

 

Figure 4.11- Photograph looking downslope from the rim of kettle site 33. A gully is visible in the
foreground and leads to a fan-like deposit downslope in the kettle bottom. Photo by T. Hobbs.

80

“naturism we

‘V

 

Figure 4.12- Photograph showing the top of the kettle bottom deposit profile at site 33. The fan
deposit shown in Fig. 4.11 is preserved as a sandy layer from 0-5 cm in this photo. This erosional
event buried the charcoal rich soil from 5-15 cm. Photo by R. Schaetzl.

related to the depth from which the sample was taken, given that there may be varying
amounts of in-mixed sand due to slopewash and bioturbation at various depths
throughout the kettle bottom deposits. To better understand the down profile textural
variation in kettles, eight kettle bottoms were depth sampled and analyzed. Such data,
explored later in this chapter, will also offer insight into the timing of Silt deposition

within kettles on the Evart Upland.

81

Supporting Evidence for an Aeolian Origin of the Kettle Bottom Silt

The textural evidence presented thus far suggests that the silt in kettle bottoms .
likely has an aeolian origin and was not winnowed out of the surrounding backslope
soils. Because the Evart Upland is dominantly composed of sandy sediments, the
spatially concentrated, silty kettle bottom deposits appear anomalous and unconforrnable.
If the kettle bottom deposits are indeed deposits composed mostly of aeolian silt, then
such a “patchy” pattern of deposition requires an important follow-up hypothesis: Kettle
bottoms are not likely the only landscape positions that have collected silt. For example,
perhaps flat, geomorphically stable surfaces should have also preserved some of the same
aeolian silt that was preserved in kettle bottoms. In contrast, steeply sloping surfaces (i.e.
backslopes) probably would not retain sigrificant amounts of aeolian silt, because they
are more prone to runoff. As a result, silt may have been retained on stable (flat)
surfaces, while silt deposited on sloping surfaces eventually got spatially concentrated in
kettle bottoms. To test whether or not this scenario applies to the Evart Upland, three
geomorphically stable sites were depth sampled to determine their silt distribution with
depth in the soil profile. The results of this additional sampling effort will offer

supporting evidence for an aeolian origin for the kettle bottom silt.

4.4 Silt Contents with Depth - Geomorphically Stable Sites

The following figures (Figs. 4.13 - 4.18) show the distribution of four textural
components with depth, at the three stable Sites sampled in the field: total silt (2-50 pm),
fine silt (12-25 pm), medium silt (25-35 pm), and coarse silt (35-50 pm). Each of these

silt components was calculated on a clayfree basis to remove the influence of

82

pedogenically derived clay. Each depth figure is preceded by a figure containing a
locator map and topogaphic map of the stable site, as well as a photogaph of the

understory vegetation.

4.4.1 Stable site-1

Stable site-1 is located on a relatively flat portion of a valley side that extends up
the southern portion of the study area (Figs. 4.13), at 382 m elevation. The total clayfree
silt content decreases from 17% to 9.5% at 75 cm depth (Fig. 4.14). The down-profile
trend of medium and coarse clayfree silt is similar, decreasing fiom ~ 4% at the surface to
~2% at 75 cm depth. However, fine clayfree silt shows a geater decrease with depth

(6% to 3% at 75 cm).

4.4.2 Stable site-2

Stable site-2 is located on a flat surface between kettles, just up-valley from
stable- site 1 (Fig. 4.15), at 391 m elevation. The total clayfree silt content decreases from
26% to 12% at 105 cm depth (Fig. 4.16). The down—profile trend of medium and coarse
clayfree silt is similar, decreasing from 5% at the surface to ~2% at 105 cm depth.
However, fine clayfree silt shows a geater decrease down-profile than medium and

coarse silt, from 10% to ~ 3.5% at 105 cm depth.

4.4.3 Stable site-3

Stable site-3 is located on a relatively flat surface surrounded by rolling ridges
upslope (to the immediate east and north) (Fig 4.17), at 406 m elevation. The total
clayfree silt content decreases from 10.5% to ~2.5% at 75 cm depth (Fig. 4.18). Similar

to the previous two sites, the down profile trend in medium and coarse clayfree silt is

83

 

 

 

Figure 4.13- Location of stable site-1 relative to the other stable site locations. B) Topographic map
showing surface characteristics of stable site-1. C) Photograph of surface vegetation at this site.

84

8 1 12141618
.l...

 

 

 

 

 

 

 

 

 

5 I...IL.LIL.-I ..
15—
25—
— total sill (2-50 microns)
. -— f silt(12-25 microns)
E35— — msilt (25-35 microns)
Q .
E * c silt (35-50 microns)
a a
9 -
a 45—
.
1
55—
65-
l
75 |lllllIIIIIIIIIIIIIIIIIIIII
0 6 8 IO 12 l4 l6 18

Percent silt (clayfree)

Figure 4.14- Depth plot showing distribution of four silt components with depth at stable site-1.

85

 

 

Figure 4.15- A) location of stable site-2 relative to the other stable site locations. B) Topographic
map showing surface characteristics of stable site-2. C) Photograph of surface vegetation at this site.

86

30

 

25~

45-

Depth (cm)

85~

 

 

 

 

 

105 l IIlrfifirl.rflllvr‘flrr.rlrrrr

0 5 10 15 20 25
Percent silt (clayfree)

 

30

 

 

total silt (2—50 microns)
f silt (12—25 microns)
m silt (25-35 microns)

c silt (35-50 microns)

 

Figure 4.16- Depthplot showing distribution of four silt components with depth at stable site-2.

87

 

 

 

 

 

 

 

 

0 100 200

Figure 4.17 A) location of stable site-3 relative to the other stable site locations. B) Topographic map
showing surface characteristics of stable site-3. C) Photograph of surface vegetation at this site.

88

L;

h—
p—

10

12

 

151

J_L

M
{It
A

1

DJ
LII

Depth (cm)
A
VI

1 L l L 1 L l l

l

 

 

 

 

 

 

‘ l

4

T‘I Tfffl

6

I I

8

I I

Percent silt (clayfree)

 

 

total silt (2-50 microns)
f silt (12-25 microns)
m silt (25-35 microns)

c silt (35-50 microns)

 

Figure 4.18- Depthplot showing distribution of four silt components with depth at stable site-3.

89

 

parallel, subtly decreasing from ~2 % to ~1% at 75 cm depth. Fine clayfiee silt decreases

more than medium or coarse silt down-profile, from ~ 4% to ~ 1% at 75 cm depth.

4.4.4 Summary of Depth Plot Data for Geomorphically Stable Sites

All three geomorphically stable sites Show an increase in clayfree silt content in
the upper profile. On average, the three profiles contain 3 ~10% up-profile increase in
clayfiee silt from the bottom to the top of the profiles (i.e. between fi'om about 80 cm to
the surface. Silt content ranges between 2-12% in the lower profiles (~80 cm) to 10.5-
26% at the surface. Fine silt, which is more readily transported by wind, increases more
than medium and coarse silt in the upper profiles of all three sites. Fine silt is also the
dominant particle size in kettle bottom deposits (recall that the silt peak of all kettle
bottom samples averaged 18 pm). These data suggest that aeolian silt has indeed been
preserved on stable surfaces, and incorporated into the soil profiles.

At the three geomorphically stable sites sampled in the field, the understory
vegetation is dominated by rice gass (Oryzopsis asperifolia). The presence of rice gass
in the understory here may be associated with slightly higher water holding capacity near
the surface, which may in turn be attributed to an increase in silt in the upper soil profile.
Indeed, rice gass is commonly found gowing on flat, sandy, upland soils, otherwise
enriched with loess, in northern Wisconsin (R. Schaetzl, personal communication, 2008).
As such, rice gass appears to be a good indicator of aeolian silt enrichment on upland,
sandy landscapes of the upper Midwest.

In other landscapes, an up-profile increase in silt content could potentially be

related to increased physical weathering, which can produce silt from sand. Mikesell et

90

 

a1. (2004) studied the degee of hornblende etching (a mineral indicator of weathering)
with depth in four sandy soils ~ 100-150 km north of the Evart Upland, to understand
how different macroclimates influence weathering intensity in soils of similar age. Their
results indicate that the intensity of up-profile hornblende etching somewhat parallels the
degee of podzolization, which in their study area, is facilitated by heavy snowfall. Their
data also show that the up-profile increase in hornblende etching is usually accompanied
by an increase in silt content. They recogiize that such silt may be due to increased
weathering in the upper profile, but also acknowledge that it may be due to an aeolian
influx. Barrett (2001) also documented silt increases in the upper profiles of sandy soils
in northern Michigan and attributed such an increase to a slight aeolian influx. Given that
soils in the Evart Upland are dry in summer and receive less snowfall than the study area
of Mikesell et al. (2004), physical weathering is not likely a significant contributing
factor to silt production in the Evart Upland. As such, the up-profile silt increase is likely
due to aeolian influx, rather than via weathering.

Unlike sloping surfaces, geomorphically stable, flat surfaces are less apt to
produce runoff. As such, flat surfaces are more likely to retain and pedogenically
incorporate aeolian silt than sloping surfaces. Indeed, data from the Buckley Flats (a flat
upland in northwest Lower Michigan) and the Grayling Fingers regions of Michigan
demonstrate that aeolian Silt is best preserved on flat, stable surfaces, and is commonly
absent on surfaces with even the slightest slope or presence of erosion (Schaetzl and
Weisenbom, 2004; Schaetzl, 2008; Schaetzl and Hook, 2008).

At the stable sites sampled, the silt increase in the upper profile is gadual and

subtly expressed within a textural backgound heavily dominated by sand. It is unclear

91

whether additions of aeolian silt at these sites were intermittent, or added during a single
depositional event, because pedogenic processes have since mixed the silt into the profile
and blurred any definitive depositional characteristics or lithologic contacts. Regardless
of the gadual nature of the up-profile silt increase, these data considered in combination
with the particle size distribution data from kettle bottoms (which illustrate that both are
composed of very similar modal silt), strongly suggests that: l) aeolian silt has indeed
been deposited across the Evart Upland after the landscape stabilized, and 2) the
degree to which silt has been preserved and incorporated into the soil may be
influenced by the slope/long-term stability of the surface, or perhaps, the paleo-
vegetation on site at the time of deposition.

Finally, the mechanism by which silt has been Spatially concentrated in kettle
bottoms is unclear, but the scenario favored herein is that perhaps, over millennia, late-
lying snowpack (which would have lasted the longest in kettle bottoms), acted as a trap
for aeolian silt, subsequently focusing it in the center-bottom portions of the kettle upon

meltout of the snow.

4.5 Spatial Patterns of Silt Content - Backsloms vs. Kettle Bottoms

The spatial pattern of silt content within backslope and kettle bottom samples
across the upland offers further evidence that the silty kettle bottom deposits are likely
the result of aeolian deposition. If the silt in kettle bottom deposits were winnowed out
of the surrounding backslopes via slopewash, then the spatial variation in kettle-bottom
silt content across the upland should be very similar for both sample types. For example,

where there is more fine silt in backslopes, there too should be more fine silt in kettle

92

bottom deposits (if slopewash were the mechanism of silt deposition in kettle bottoms).
To compare the spatial variation of silt content between both sample types, the textural
data were linked with the sample locations in ArcMap, and gids of three ranges of silt
content (coarse silt (3 5-50 pm), medium silt (25-35 pm), and fine silt (12-25 um)) were
created using an ordinary kriging interpolation method in Geostatistical Analyst
(ArcMap, Redlands, CA).

The results of the kriging operation revealed an increasing discordance between
the two sample types with decreasing silt particle size (Fig. 4.19). For example,
backslopes exhibit similar spatial patterns for all three silt components, having geater
concentrations of all three silt ranges in the east and north (represented by darker brown)
(Fig 4.19). Because the sediments comprising backslope samples were deposited in a
glaciofluvial setting, the spatial variation in coarse, medium, and fine silt content should
be (and indeed, are) very similar because, in the context of transport and deposition,
flowing water did not crisply discriminate between coarse and fine silt. In contrast, kettle
bottoms exhibit different spatial patterns for all three silt components, having geater
concentrations of fine silt in the southeast and central portions of the study area than the
rest of the study area (Fig. 4.19). The discordance between the two sample types would
be expected if fine silt has been deposited across the upland by a process that is
landscape-scale, like wind, rather than a small scale, Slope-based process like Slopewash.
The different spatial distribution of fine silt between kettle bottoms and backslopes
further supports the conclusion that the majority of silt in kettle bottoms is the result of

aeolian deposition.

93

acne—39a: 5:: En one... new 9::— mEotea 238—

8:32— onb 2:53 :23 .3.— Eonohme one in .533 .3 82a.» 2: £9533— .393 2:53 52— 5 3.8 non—nae 29:23.55 an: 18:39.3.— E 33:3
in .5320 do...“ he...» 2: 232. 25:32 29:: E895?— Beu Jew—m $2..» 2: .«o =a E a: a 5.52 .95.. 3335 BEE: E323 2:8. E... 95..
ne: 335a... one—fleas 59252. GAME Ev. on: “Ea Ana—.1593 25:3:— Afio—v in 8:3 "35.5955 Ev. 3...: no at?» near:— Eai—EO 426 95»:—

 

AE: mNINC Em 0:: -wEosom AS: mn-va Em E285 -wEouom AS: 8.9% Em wflwoo -wEouom

 

 

:5 mum: so one -ooaoaxoom E5 8an an 528:. -ooageeom 25 8-3 an 8:8 -ooaoaxosm

 

 

94

WWW

Analysis of the kettle bottom deposits revealed significant textural variation,
particularly with respect to sand content (Fig. 4.4), suggesting that they may be texturally
heterogeneous with depth. As suggested earlier, variation in sand content in kettle
bottom deposits is probably due to influxes of slopewash sand. Eight of the sampled
kettle bottoms had evidence of buried soils within the kettle bottom silt deposits, during
the initial field sampling effort. The presence of buried soils in the kettle bottom
deposits, coupled with the varying amounts of sand with depth, indicates that surfaces in
this landscape have experienced periods of stability and instability in the past. Surface
instability could potentially lead to different kinds of sediment deposition (aeolian silt or
slopewash sand) in kettle bottoms, further explaining why kettle bottom deposits are
texturally heterogeneous with depth. Therefore, the eight kettle bottoms containing
buried soils were depth sampling to 1) better understand the depositional history of silt
and sand in kettle bottoms, and 2) determine the degree to which the kettle bottoms are
texturally heterogeneous with depth. Such evidence will be used to confirm or deny that
the textural variation among kettle bottom deposit samples (as revealed in Fig 4.4) is

related to the depth from which the sample was taken.

95

4.6.1 Explanation of depth-plot method and depth-plot figures

4.6.1a Depth-plot method

The method used in this thesis to display down profile changes in soil texture in
this section is based on a method developed by Beierle et a1. (2002), originally applied to
lake sediment cores. Their method is used here because it visually enhances subtle
down-profile changes in particle size distribution, and therefore, highlights depositional
processes that may otherwise remain unseen in standard depth-plots of simple statistical
measurements, such as mean particle size. By incorporating the entire particle size
distribution at every sampled interval into a continuous grid surface, subtle depositional
events in the kettle bottom deposits are made more evident.

The graphs use three axes of data (x, y, z) for each depth-sampled interval. The
x-axis corresponds to the particle size, y-axis to the depth of sample, and z-axis to the
volume percent of the sample (obtained from the near-continuous particle size
distribution curve data from the Malvem Mastersizer; see Methods chapter). Essentially,
the graphs display the particle size distribution curve at each depth-sampled interval, and
interpolate z-axis values (volume percent) between sampled intervals. The result is a
continuous grid surface of particle size distribution characteristics throughout the entire

soil profile.

4.6.lb Explanation of depth plot figures

The following figures (Figs. 4.20 - 4.35) display data from eight depth-sampled

kettle bottom deposits. Each of the figures are accompanied by a facing page that

96

contains a locator map and a topographic map, which show the location of the eight
sample sites relative to each other, and a close up of the surrounding topography and
kettle shape, respectively. In addition to the geography of the depth sampled kettles,
some basic morphometric characteristics of the kettles themselves and the surrounding
area are also provided. These characteristics include: elevation, kettle area, drainage
basin area, aspect of backslope sample, and backslope gradient. Kettle area and drainage
basin area were previously calculated in ArcMap (Redlands, CA). Aspect and gradient
data were determined in the field.

Each kettle that was depth sampled contains at least one buried soil. In each of
the depth plot figures that follow, a solid black line marks the depth of the top of a buried
soil. Coarse, sandy substrate sediments lie beneath the kettle bottom deposits in all of the
kettle bottoms that were depth-sampled. A dashed line on the depth plot figures marks
the boundary (lithologic discontinuity) between the coarse sandy substrate sediments and

the overlying silty sediments.

97

 

 

   

 

M
Elevation: 378 m
Kettle area: 21,300 m.sq.
Drainage basin area: 266,100 m.sq.

Aspect of backslope sample: NW

 

 

. Backslope gradient: 26%

 

      

to 75 150

 

ml
300

Figure 4.20— A) Location of site 17 in relation to other depth sampled kettles on the Evart Upland. B)
Topographic map showing kettle site 17 with additional basin characteristics (right).

98

Site 17

 

30

50

70

Depth (cm)
o
o

 

 

 

 

130 Volume %

 

150 . l . . t
l 10 100 1000
Particle Size (microns)

Figure 4.21- Depositional sequence of kettle bottom sediments at site 17. The dashed line marks the
top of the coarse, sandy substrate sediments. The solid line marks the top of a buried soil.

99

 

 

    

 

S—ite 24

      

i levation: 372m

ettle area: 175 m.sq.

!

(«E

F)

rainage basin area: 84,457 m.sq.

pect of backslope sample: E

ackslo radient: 45%

 

 

   

Figure 4.22— A) Location of site 24 in relation to other depth sampled kettles on the Evart Upland. B)
Topographic map showing kettle site 24 with additional basin characteristics (right).

100

Site 24

 

20

40

60

80

.—
O
O

Depth (cm)
5
o

 

 

 

 

 

140
160
180
Volume %
200
220 . - , a
l 10 100 1000

Particle Size (microns)

Figure 4.23- Depositional sequence of kettle bottom sediments at site 24. The dashed line marks the
top of the coarse, sandy substrate sediments. The solid line marks the top of a buried soil.

101

 

 

 

 

Site 33

   
 
  

Elevation: 396 m
Kettle area: 18.254 m.sq.
Drainage basin area: 185,820 m.sq.

Aspect of backslope sample: WNW

 

 

Backslope gradient: 30%

 

 

   

--:—m
ho 75 150 300
l I \I i J‘

 

Figure 4.24- A) Location of site 33 with in relation to other depth sampled kettles on the Evart
Upland. B) Topographic map showing kettle site 33 with additional basin characteristics (right).

102

EM

 

 

 

 

 

14
13
12
ll
10
9
8
A 7
5 6
5 5
3
G 4
3
2
1
0
Volume %

 

140

1 10 100 1000
Particle Size (microns)

Figure 4.25- Depositional sequence of kettle bottom sediments at site 33. The dashed line marks the
top of the coarse, sandy substrate sediments. The solid lines mark the top of buried soils.

103

 

 

    

 

Site 42

 

Elevation: 402 m
‘ Kettle area: 2,624 m.sq.
Drainage basin area: 29,700 m.sq.

Aspect of backslope sample: N

 

 

lBackslope gradient: 18%

 

Figure 4.26- A) Location of site 42 with in relation to other depth sampled kettles on the Evart
Upland. B) Topographic map showing kettle site 42 with additional basin characteristics (right).

104

Site 42

 

 

Depth (cm)

 

 

 

 

Volume %

195

1 10 l 00 1000
Particle Size (microns)

Figure 4.27- Depositional sequence of kettle bottom sediments at site 42. The dashed line marks the
top of the coarse, sandy substrate sediments. The solid line marks the top of a buried soil.

105

 

 

 

 

 

M

‘ Elevation: 378 m

Kettle area: 26.579 m.sq.

Drainage basin area: 331,013 m.sq.

Aspect of backslope sample: NE

 

lBackslope gradient: 42%

 

 

  

 

 

Lo -7:—5 150

ll

 

Figure 4.28- A) Location of site 45 with in relation to other depth sampled kettles on the Evart
Upland. B) Topographic map showing kettle site 45 with additional basin characteristics (right).

106

Site 45

 

 

Depth (cm)

 

 

 

 

Volume %

110

l 10 100 1000
Particle Size (microns)

Figure 4.29— Depositional sequence of kettle bottom sediments at site 45. The dashed line marks the
top of the coarse, sandy substrate sediments. The solid line marks the top of a buried soil.

107

 

 

    

 

Site 49

 

Elevation: 387 m
Kettle area: 72,850 m.sq.
Drainage basin area: 180,434 m.sq.

Aspect of backslope sample: W

 

f Backslope gradient: 29%

 

Figure 4.30- A) Location of site 49 with in relation to other depth sampled kettles on the Evart
Upland. B) Topographic map showing kettle site 49 with additional basin characteristics (right).

108

Site 49

 

 

 

 

 

 

14
13
12
ll
10
9
8
fl‘ 7
g 6
5 5
8' 4
CI
3
2
l
0
Vdume%

l 10 100 1000
Particle Size (microns)

Figure 4.31- Depositional sequence of kettle bottom sediments at site 49. The dashed line marks the
top of the coarse, sandy substrate sediments. The solid lines mark the top of buried soils.

109

 

 

   

 

M
Elevation: 399 m
Kettle area: 9,438 m.sq.
Drainage basin area: 51,933 m.sq.

Aspect of backslope sample: E

 

 

Backslope gradient: 20%

 

    

203'1 I

Figure 4.32— A) Location of site 52 with in relation to other depth sampled kettles on the Evart
Upland. B) Topographic map showing kettle site 52 with additional basin characteristics (right).

'0 so 100

110

Siiil

 

20

40

6O

00
O

Depth (em)
8
o

120

 

 

 

 

140 Volume %

160

 

1 70 .
1 10 100 1000
Particle Size (microns)

Figure 4.33- Depositional sequence of kettle bottom sediments at site 52. The dashed line marks the
top of the coarse, sandy substrate sediments. The solid line marks the top of a buried soil.

111

 

 

 

 

  
   

 

Site 53

 

Elevation: 411 m

 

 

Backslope gradient: 52%

 

j 0%

0 50 100 _200 "1K1 /()/ k ‘N-n

Figure 4.34- A) Location of site 53 with in relation to other depth sampled kettles on the Evart
Upland. B) Topographic map showing kettle site 53 with additional basin characteristics (right).

112

SM

 

n—p—nl—II—II—n
OHNUJ-b

Depth(cm)
o—NMAUIOKQOOO

 

 

 

 

Volume %

 

1 10 100 1000
Particle Size (microns)

Figure 4.35- Depositional sequence of kettle bottom sediments at site 53. The dashed line marks the
top of the coarse, sandy substrate sediments. The solid lines mark the top of buried soils.

113

4.6.2 Integrating the Data in Depth-Plot Figures

Figures 4.20 - 4.35 illustrate that each kettle bottom contains a unique
depositional sequence and different numbers of buried soils. In some kettles, intervals of
sandy sediment abruptly alternate with intervals of silty sediment (Fig. 4.23), while in
other kettles, sediments in the profile are well mixed, and no clear mode of deposition is
evident (Fig. 4.31). These differences indicate that the rates of sand deposition, silt
deposition, and pedogenic upbuilding are different, and to a degree, site-specific, for each
kettle.

The strength with which depositional intervals are expressed may be dependent
upon the balance between depositional processes (i.e. slopewash) and pedogenic
processes (i.e. bioturbation and pedogenic upbuilding) affecting the kettle bottom. This
balance may be controlled by the morphometric characteristics of the kettle basin. For
example, small, steeply sloped kettles draining a large area may be more influenced by
slopewash processes than pedogenic processes because 1) as backslope gradient
increases, runoff and erosion are more likely, and 2) small kettles that drain a large area
can potentially integrate and “focus” more overland flow than large kettles that drain a
small area. Therefore, the depositional sequence in small, steep kettles may have more
clearly preserved sandy intervals in the kettle bottom than large, shallow kettles, because
sediment deposition would be more frequent, and perhaps overwhelm the rate of
pedogenic upbuilding in the former.

To determine if this interpretation applies to the eight depth-sampled kettles,
simplified, integrated versions of the depth plot figures were created and compared

relative to each other (Fig. 4.36). Each kettle site was ranked in order from steepest to

114

shallowest backslope, then again by kettle area-to-drainage basin ratio. The depth-plots
are grayscale shaded based on particle size distribution, and range between two end
member depositional processes: aeolian silt or sandy influx. Aeolian silt intervals are
indicated by a unimodal particle size distribution with the modal grain size in the fine-
medium silt range (~10-35 um). Sandy influx intervals are indicated by a unimodal
particle size distribution with the modal grain size in the sand range (generally coarser
than 100 pm). Mixed intervals, which fall between the two end members on the scale,
are indicated by a bimodal particle size distribution with no clear dominance by either
silty or sandy sediments. These units likely represent periods of pedogenic upbuilding
when neither depositional process (aeolian silt deposition or slopewash sand) dominated,
or perhaps are the result of post-depositional mixing (bioturbation). Mixed intervals can
be further subdivided into coarsening upward or fining upward sequences depending on
the up-profile trajectory of the PSD curves in the depth plots.

In effect, if the above assumptions are true regarding the role of basin
characteristics in controlling the balance between slopewash processes and pedogenic
processes, then the profiles should become more texturally homo genous from left to right
in Figure 4.36, reflecting a decreasing influence by slopewash processes, and increasing
capacity to pedogenically incorporate sediments. Qualitatively, the depth plots in Fig.
4.36 do appear to be slightly more homogenous from left to right. For example, site 42
(far right) contains a mixed profile and fewer distinct depositional intervals than site 33
(middle left). Furthermore, site 53 (far left) has three buried soils, whereas site 42 has
one buried soil, suggesting that slopewash processes are more common in the former.

Such comparisons between sites suggest that the above assumptions regarding the role of

115

can... :2... e. 9520.. 2.8... o. .55.... 2... 9... 23.3583... 2:9... 82.... 5:3... 2....
En...— uwa=_a._..-...-au..a .8”...— u.... mono—8.9.... .8323... e. 32. 3:... Eu...— ouaiahuéfiau: .3525 E... meme—32:. $2.33 ”none—8.2... w..:...=.....=.w
e... E9... 5.52.2. £2.33...“ 3 3:33.38... ”523.5... he 5...... E ens—E... .mné - cu... 83w: 5 82.3.9... 0... .«e 2.239. .55155 -3”... 9:5...—

 

 

 

 

 

 

 

Qqacgov 0.388%
cosfioaov ecum-fia_oofi.8 25-33.38 conaomuv 35.3.5
32...; «.822 «.952 “am canon... 3.5m
mom potsm . . .. ,. H
03.... Eman— oWwEwavéfuoua 053— wfimwouoofl «GOESm one—mV—OMD mammaosofi

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
   
 

 

 

 

 

 

 

 

 

 

 

 

 

. N... “am . mm an 9. em a ”em 9. 2a mm an em 2a m new

A _

“goaomofioa wages...

 

 

116

basin characteristics in controlling the strength with which depositional intervals are
expressed may have a slight influence on the depositional sequence in kettle bottoms.
Other factors have probably influenced the depositional sequence in kettle
bottoms in this landscape, such as paleo-vegetation changes (Almendinger and Hobbie,
1992) and historical precipitation (Lamoureux, 2000). As of yet, the influences of these
factors remain ambiguous in the context of this landscape, and are not evident in the data
presented herein. However, what can be said of the data in Figs. 4.20 - 4.35 is that
depositional events cannot be correlated between the eight sampled kettle basins.
Therefore, the deposition of silt and sand has probably occurred differently in each kettle,
suggesting that surface instability and subsequent redistribution of sediments into
depressions has been episodic and localized. Overall, the depth plot data suggest that this
landscape is highly sensitive to disturbance, which is reflected in the sequence of

sediments within kettle bottoms.

4.6.3 Summary of Depth-Plot Figures

The depth plots in Figs. 4-20 - 4.35 clearly show that the texture of kettle bottom
deposits fluctuates with depth. Each kettle bottom deposit has a unique depositional
sequence that fluctuates between two end members of deposition, silt (aeolian deposition)
and sand (slopewash deposition). In many cases, intervals in the kettle bottom sediments
are mixed with respect to particle size distribution, suggesting no clear dominance by
either depositional process. Fining upward and coarsening upward mixed sequences may
indicate pedogenic upbuilding in response to ever increasing aeolian deposition, or

increasing sand deposition, respectively. Mixed intervals may be the result of varying

117

degrees of bioturbation. Furthermore, each kettle bottom contains a unique sequence of
buried soils, reflecting former periods of landscape stability and instability. The different
number of buried soils in each kettle implies that disturbances leading to the deposition of
sediments were localized, rather than affecting the entire study area. Furthermore, the
preservation of such disturbance intervals in the kettle bottom deposits may be related to
the balance between depositional processes and pedogenic processes, which are
ultimately influenced by the morphometry of the kettle basins. In general, the data in
Figs. 4.20 - 4.35 indicate that this landscape is highly sensitive to disturbances, as
reflected in the sequence of sediments in the kettle bottoms. '

As suggested earlier, fire may be the disturbance mechanism responsible for
facilitating the redistribution of sediments in this highly sensitive landscape. Indeed,
charcoal was found in many of the buried soils that were revealed during the depth
sampling effort. The charcoal found in these soils provided an opportunity to
radiocarbon date the timing of such disturbances. The following section will explore
evidence from buried soils in the study area, which revealed details about the timing of

aeolian silt deposition across the Evart Upland.

118

4.7 Timing of Silt Deposition

4.7.1 Evidence from Site 53

During the depth sampling effort, a buried soil was discovered at site 53, which
had developed within the sandy, substrate sediments that underlie the sequence of silt
deposition (Fig. 4.37). Though such evidence offers no absolute date on the timing of silt
deposition, the presence of a strongly developed, relatively thick (~15 cm), buried A-
horizon within the sandy substrate sediments at this site suggests that there was a
significant period of landscape stability (and plant growth) in the post-glacial period that

pre-dated silt deposition in the study area.

4.7.2 Evidence from Site 33

Charcoal is abundant within buried A horizons in the kettle bottom profile at site
33 (Fig. 4.38). Approximately 0.2 grams of charcoal fragments (<O.5 cm in size) were
gathered from the lowermost buried A horizon (~120 cm depth) and radiocarbon dated
together as one sample, using AMS to determine the minimum limiting age of silt
deposition in this kettle (Fig. 4.39) (see Methods). In addition to the charcoal fragments,
many small twig-like fragments and seeds were extracted from the buried soil. These
fragments could not be identified to species, due to their small size. Furthermore,
amorphous black-colored silica fragments were found in the sandy matrix after sieving
(Fig. 4.40). These fragments are interpreted to be sand grains that were partially melted
and re-crystallized due to fire. The charcoal fragments yielded an age of 6,840 :1: 30 (1 a)
years BP. The very small error range, despite the many fragments used to obtain the

date, strongly suggests that the charcoal fragments are the result of only one fire event.

119

 

 

Figure 4.37- This photograph shows a trough containing the sequence of kettle bottom sediments at
site 53. The pale tan sediments at the bottom of the photo are the sandy substrate sediments
(lowermost C-horizon). Overlying the buried soil is the first interval of silty, aeolian sediment
(brown colored). Photo by R. Schaetzl.

120

 

       

Site 33
20 1

Depth (cm)

 

 

l l0 100 I000
Particle Size (microns)

Figure 4.38~ A) Photo and data from the soil profile excavated at site 33. This site contains four
buried soils, with charcoal randomly spread throughout the profile, but mostly concentrated within
the soils. B) Depth plot showing the down profile textural variation at site 33. Black lines indicate
the top of buried soils. C) Close up of the lowermost buried soil from which charcoal was extracted
at ~ 120 cm. Note: the one-meter mark on the tape measure is above the extent of the picture,
therefore, the 20 cm mark in this frame is actually 120 cm. Photos by R. Schaetzl.

121

 

Figure 4.39— A) and B) Woody charcoal pieces commonly found in buried soil at site 33. C) Seeds
were also found in the buried soil. D) Many small fragments of charcoal sticks (~0.25-0.5 mm) within
the matrix of sand grains.

122

 

    

" as ' n
..‘

v {’1 V. ‘.

 

Figure 4.40- Pieces of amorphous, black-colored silica found interspersed within the lowermost

buried soil at Site 33.

123

 

4.7.3 Evidence from Site 17

During the depth-sampling procedure at site 17, a number of charcoal pieces > 0.5
cm in diameter were found at ~120 cm depth, slightly below the lowermost (and only)
buried soil in the kettle bottom profile. These pieces were carefully extracted from the
bucket auger and radiocarbon dated, yielding and age of 920 :l: 20 (1 a) years BP. The
small errors associated with the two radiocarbon dates reported above suggest that the
charcoal found within these buried soils each resulted from individual fire events.
Furthermore, the buried soil at site 17 formed within a silty interval, and was later
covered by more silt (Fig. 4.20). The stratigraphic position of aeolian silt over and under
the charcoal-rich buried soil at site 17 suggests that silt deposition was contemporaneous

with, or at least bracketed in time, a fire event or events.

4.7.4 Summary of Evidence for Timing of Silt Deposition on the Evart Upland

The data presented herein suggest that aeolian silt deposition on the Evart Upland
has occurred episodically. The lowermost buried soil at site 53, which formed in the
sandy substrate sediments, indicates that the Evart Upland experienced a period of
stability in the post-glacial period that pre-dates at least some of the aeolian silt
depositional period(s). Furthermore, the two radiocarbon dates from site 33 and site 17
yielded ages of 6,840 i 30 BP and 920 3: 20 years BP, respectively, indicating that fires
occurred in this landscape in the middle and late Holocene. The charcoal fragments used
to obtain these dates were gathered at or below soils that are buried by sediments
indicative of drastic changes in depositional process. As such, the charcoal fi'agments

record distinct fire events that may have caused localized surface instability, thereby

124

 

 

facilitating sedimentation (either aeolian silt or slopewash sand) in kettles. Overall, these
data suggest that the Evart Upland is an unstable, geomorphically sensitive landscape that
has experienced various periods of stability and disturbance throughout the Holocene. To
summarize, three important conclusions can be made regarding silt deposition on the
Evart Upland:
1. Aeolian silt deposition was episodic, and spatially discontinuous (i.e.
not all kettles were collecting silt at the same time)
2. Some silt deposition events did not begin until well after the landscape
had stabilized, and
3. Episodic silt deposition continued to occur well into the middle and
late Holocene.

The evidence presented thus far indicates that the majority of the silt in kettle
bottoms is loess that was deposited episodically during the Holocene, and is associated
with local fire disturbances. Furthermore, this loess is unequally distributed across the
upland, being more concentrated in lower, stable landscape positions (i.e. kettle bottoms
and flat sites) than higher unstable landscape positions. The uneven and patchy
distribution of loess across the upland suggests that perhaps it is locally redistributed silt
that was winnowed out of surfaces disturbed by fire. To test this interpretation, silt
within kettle bottom deposits and adjacent backslopes samples were extracted and
compared using X-ray diffraction techniques (see Methods). The result of this analysis

follows.

125

4.8 Source of Kettle Bottom Silt - Local vs. Extra Regional?

To better understand the possible source(s) of loess comprising kettle bottom
deposits, the silt mineralogy of a selection of kettle bottom/backslope pairs was
determined using X-ray diffraction. Kuffinan (2003) used a similar approach in
comparing mineralogical differences between presumed aeolian silt and local silt, in an
investigation into the origin of high mountain loess deposits in the German Alps. If
significant mineralogical differences occur between backslope silt and kettle bottom
deposit silt, then it is reasonable to suggest that the kettle bottom silt is from outside the
region (allochthonous dust). On the other hand, if no (or minimal) significant
mineralogical differences are apparent between the two sample types, it is reasonable to
assume that l) the silt in kettle bottoms is more local in origin, or 2) the source region has

the same mineralogical composition.

4.8.1 Evidence from Texturally Different Kettle Bottom/Backslope Pairs

Four pairs of samples were chosen for the silt mineralogy comparison. The kettle
bottom deposit samples in each pair have as much as ~51% more clayfree silt than their
adjacent backslopes (Figs 4.41 and 4.42). Therefore, mineralogical differences (if
evident) would most likely be present in these kettle bottom-backslope pairs than others
(if the silt were allochthonous) because these kettle bottom samples would have had the
least influx of sediment from the surrounding backslopes.

The first two of these pairs were obtained from kettle sites that exhibit the greatest

difference in silt content (of all kettles sampled) between the kettle bottom deposit

126

 

sample and kettle backslope sample (sites 28 and 31) (Fig. 4.41). The most common
mineral found in these two kettle bottom-backslope pairs is quartz. Two other minerals
of lesser intensity are also found: K-feldspar, and plagioclase feldspar (Fig. 4.43). The
diffraction patterns in Fig. 4.43 suggest that the silt in backslope samples and kettle
bottom samples are mineralogically similar (i.e. there are no mineral peaks that are
unmatched among kettle bottom and backslope pairs), and are from the same population,
or source sediment. As mentioned previously, it is possible that the source sediment is
extra-regional, but has the same mineralogical composition as the silt deposited on the
upland. However, given that the silt in kettle bottoms was deposited contemporaneously
with macroscopic charcoal (i.e. locally derived charcoal), it is unlikely that the silt is
extra-regional. As such, the kettle bottom silt in these samples is likely locally

redistributed loess.

4.8.2 Evidence from Intervals of Aeolian Deposition in Depth-Plots

The second pair of kettle bottom/backslope samples that was chosen for
mineralogical comparison contains discrete intervals of aeolian silt deposition in the
kettle bottom profile, based on textural evidence from the depth plots (Fig. 4.42). These
sample pairs are from sites 24 and 42. The two aeolian intervals are indicated in the
depth plot by a unimodal particle size distribution with the modal grain size in the silt
range (~lO-25 pm). For consistency with regard to the previous samples, the 20-53 pm
range of silt from these extracted and compared to the same silt range in the adjacent
backslopes. The aeolian intervals within the kettle bottom deposits at sites 24 and 42

appear to be mineralogiCally similar to their adjacent backslope samples (F ig.4.44).

127

 

 

 

Again, there are no mineral peaks that are unmatched among aeolian intervals in the
kettle bottom and backslope pairs. As such, the intervals of aeolian silt deposition are

most likely comprised of local silt.

 

128

H3

 

    
   

 

 

 

 

 

 

8%
_ 20-53
I microns
7‘.
63
1
\° .
O
u 5‘.
E .
2 I
O .4
> 4 :
3%
2%
1%
n ‘ .....w-""""'mm"”mm” I
u r l . r . l l n V 11 T . r r V] . l r r v r .
0.5 l 10 100 1000
Particle size (microns)
— Bottom 28
------- Backslope 28
— Bottom 31
------- Backslope3l

 

 

Figure 4.41- Particle size distribution curves for two of the most texturally different kettle bottom-
backslope pairs. The range of silt-sized particles extracted from both is outlined (20-53 microns).

129

 

E — Bottom 24 (150 cm)
_: ------- Backslope 24

7—_ —- Bottom 42 (70 cm)
------ Backslope 42

   

_,.--: ........
x: ......
“nu""I"
V I

20-53
pm

 

\"
.

-
J‘,

 

 

 

 

VI'TY‘I rvvrlrl

0.5 I 10

 

r rrrvrv‘

100 1000

Particle size (microns)

Site 24

Depth (cm)
I 8 o
O C O

_.
00
O

  

220

Site 42

    

35

\1
U1

Depth (em)
3

.—
LII
LI]

195

l 10 100 1000 1 10 100 1000

Particle Size (microns)

Particle Size (microns)

Figure 4.42- Bottom: sites 24 and 42 both have distinct intervals of aeolian deposition in the kettle
bottom as indicated in the depthplots (intervals circled). Above: Particle size distribution curves of
the two aeolian intervals and their adjacent backslopes. The 20—53 pm fraction of these samples and

their adjacent kettle backslope samples were

extracted for X-ray diffraction analysis.

130

 

 

Relative counts

 

 

 

 

 

 

 

 

 

 

 

 

 

Degrees 2-theta

 

....... 28 Backslope
__ 31 Bottom

------- 31 Backslope

 

 

 

Figure 4.43- Results of the X-ray diffraction analysis for the two most texturally different kettle
bottom-backslope pairs. Each kettle bottom-backslope pair is color-coded. Dashed lines represent
backslope samples, while solid lines represent kettle bottom samples. Each diffraction pattern is
offset 500 counts above the underlying pattern for ease of visual comparison. As such, the Y-axis
should not be read as an absolute count. Three significant minerals are present in the samples:
Quartz, K—feldspar, and plagioclase feldspar.

131

 

 

 

7000 -_

6000

I‘llllllll

5000

Relative Counts
A
O
O
1°

 

 

 

 

.Iilxllll1

t

h—‘
C
O
O
J

 

 

 

 

O

25 26 27 28 29
Degrees 2-theta

 

— Bottom 24 (150 cm)
....... Backslope 24

— Bottom 42 (70 cm)

 

 

....... Backslope 42

 

Figure 4.44— Results of the X-ray diffraction analysis for the two kettle bottoms containing distinct
aeolian units. Each kettle bottom-backslope pair is color-coded. Dashed lines represent backslope
samples, while solid lines represent kettle bottom samples. Each diffraction pattern is offset 500
counts above the underlying pattern for ease of visual comparison. As such, the Y-axis should not be
read as an absolute count. Three significant minerals are present in the samples: Quartz, K-feldspar,
and plagioclase feldspar.

132

4.8.3 Summary of X-ray Diffraction Analysis

The silt in kettle bottom deposits was most likely locally derived. The diffraction
patterns of silt in the two selected kettle bottom/backslope pairs appear to be
mineralogically similar. Quartz is by far the most abundant mineral in the eight samples
analyzed, and indeed probably the most abundant mineral found within soils of the
surrounding regions, which, in this part of northern Lower Michigan, are dominantly
comprised of outwash sands (Schaetzl and F orman, 2008).

The intensity of quartz in the diffraction patterns may have “drowned out” the
presence of other less significant minerals that could potentially show differences
between kettle bottoms and backslopes. As such, it may have been helpful to combine
data from X-ray diffraction analysis with data on heavy minerals within the two sample
types (Kufmann, 2003). However, the diffraction patterns nonetheless illustrate that there
are no significant mineral peaks present in one sample that are not present in another.
Therefore, it is likely that the silt in kettle bottoms is locally redistributed loess that was
winnowed out of nearby exposed surfaces and deposited across the Evart Upland, rather

than being derived from some extra-regional source area.

133

 

4.9 Origin of Kettle Bottom Deposits - Summary of Evidence

Table 4.3 summarizes the evidence regarding the characteristics and origin of the
silty kettle bottom deposits in the Evart Upland. The kettle bottom deposits are
dominated by aeolian silt. This silt was probably winnowed out of nearby disturbed
surfaces either contemporaneously or immediately following fire. The charcoal used to
date the fire events that are responsible for the destabilization of surfaces in and around
the Evart Upland indicate that such events occurred during the Holocene. Fires in this
landscape were probably not spatially extensive, given that the upland is topographically
complex, and hence, fires probably did not spread easily. Perhaps kettles nearest a
surface disturbed by fire collected more silt than kettles further from the source. Such an
interpretation is supported by the fact that 1) each kettle has collected a different amount
of silt, and 2) the sequence of sediments and buried soils within each of the depth

sampled kettle bottom deposits are not correlative across the upland (Figs. 4.20 - 4.35).

134

 

Table 4.3- Summary of evidence regarding the origin of kettle bottom deposits.

The silt comprising kettle bottom deposits.

 

Characteristic

 

Evidence

 

has an aeolian origig

 

The kettle bottom deposits...

 

 

a) lie unconformably over the underlying sandy
substrate sediments.

b) have abrupt lateral boundary with the surrounding
backslope sediments (i.e. were not winnowed out of
the surrounding slopes via slopewash).

c) contain 44% more clayfree silt than backslopes.

d) have a modal silt grain size of 18 pm (as
compared to 45 pm in the backslopes).

 

Other evidence

 

Ia) Nearby flat, stable surfaces have a 10% up profile
increase in clayfi'ee silt suggesting subaerial
deposition via eolian processes.

b) The spatial distribution of fine silt in kettle
bottoms is different than that in backslopes,
suggesting significant sorting via wind (rather than
slopewash)

 

 

was deposited episodically throughout the
Holocene

a) Charcoal found in buried soils provided
radiocarbon ages of ~ 6840 and 920 years BP

 

b) Charcoal is commonly found in other buried soils
in kettle bottoms (not yet radiocarbon dated)

 

was winnowed from local sediments

The diffraction patterns of silt in representative
kettle bottom and backslope samples are
qualitatively the same

 

was deposited contemporaneously with £113

Depthplots (Figs. 4-17 - 4.25) show that aeolian silt
often buries and/or contains charcoal fragments

 

 

1

35

CHAPTER 5 - Paleoenvironmental Significance

5.1 Soil Charcoal as an Indicator of Fire

The charcoal in the kettle bottom deposits has utility as dating tool, as shown
above; it also can be used as a proxy for reconstructing local historic fire regimes (Berli
et al., 1994; Gavin et al., 2003). However, using soil charcoal as an indicator of former
fire regimes can be problematic due to the variability in processes related to transport,
burial, and pedogenic mixing, as well as the methods used to assess the age of the
charcoal. As such, many factors must be considered before making confident
interpretations regarding the intensity and periodicity of former fires.

In the context of transport and deposition, there are two main types of charcoal:
primary and secondary. Primary charcoal is deposited directly from a fire, or shortly
thereafter, whereas secondary charcoal is deposited much later, sometimes transported
great distances by wind or water before it reaches its final resting place (such as a soil or
nearby lake basin) (Whitlock and Larsen, 2001). In general, the size of charcoal pieces
usually decreases with distance from the source of the fire; much of it is carried alofi by
convective air currents (Pisaric, 2002), or fragmented if traveled a great distance over
land via runoff or slopewash. After it has been deposited and buried, various aspects of
pedogenesis, such as bioturbation and freeze-thaw processes, may translocate and mix the
charcoal within the soil profile. In general, Gavin (2003) has found that, because
deposition of charcoal occurs at the surface, charcoal within lower mineral horizons tends
to be much older than charcoal found in organic horizons, as would be expected due to
soil mixing processes. In cases when the soil charcoal is not in situ (or in other words,

eroded and re—deposited), it no longer provides an accurate age estimate for the horizon in

136

which it is found. Such was the case described by Arbogast and Packman (2004) for a ~
39 ka radiocarbon age estimate on charcoal found at a depth of ~1 .5 m in a Holocene
inland dune field in northern Lower Michigan.

The radiocarbon-dated charcoal from the kettle bottom deposits in the Evart
Upland likely resulted from local fires. A number of charcoal pieces found immediately
below the buried soil at site 17 were >0.5 cm. As such, they are most likely primary
charcoal, indicating that they have not been transported a great distance, and are,
therefore, byproducts of a local fire. On the other hand, the charcoal pieces found in the
lowermost buried soil at site 33 were <05 m. It is unclear whether the size of these
charcoal pieces is related to long distance transport prior to burial, or continual post-
depositional fragmentation in the >6000 years since they were buried. However, both
radiocarbon dates have small errors (20-30 years) despite the fact that multiple small
pieces of charcoal were used to obtain each radiocarbon date. As such, these charcoal
pieces likely resulted from single fire events, and therefore, the sediments deposited
within kettle bottoms thereafter are probably associated with the individual fire events.

Despite the number of problems associated with using charcoal as a fire proxy,
the presence of charcoal in a soil horizon nonetheless clearly documents the former
occurrence of fire in the area. The two radiocarbon dates of charcoal associated with
buried soils in the study area have very different dates (6820 i 30 and 920 i 20 years
BP). When viewed in conjunction with the presence of other charcoal-rich buried soils
found in kettle bottoms (that are yet undated), the data suggest that fire disturbances
occurred episodically in this region from at least the middle through the late Holocene,

and perhaps at all times since final landscape stabilization ca 13-15 ka.

137

5.2 Fire on Sandy Substrates

Many workers have demonstrated that dry, sandy outwash or ice-contact
landforms in Michigan primarily supported fire-prone species prior to European
settlement (Brubaker, 1975; Whitney, 1986; Leahy and Pregitzer, 2003). Perhaps the
most fire-prone of these landform/vegetation associations are the various sandy outwash
plains in northern Michigan that now support Jack pine forest (Pinus banksiana) (Simard
and Blank, 1982). Host and Pregitzer’s (1992) analysis of current species distributions
on the sand-dominated glaciofluvial landforms of northwestern Lower Michigan
suggested that variation with respect to landform has greatly influenced the successional
pathways of forest development, which in turn influences the likelihood of fire. For
example, in the Upper Peninsula of Michigan, Brubaker (1975) found that Holocene
vegetation succession varied significantly with respect to substrate type. The flat, dry,
sandy outwash deposits of the Yellow Dog Plains facilitated the development of jack pine
forest throughout the Holocene and into pre-settlement times, while more mesic, finer
textured till deposits in the surrounding areas tended to support eastern white pine.
Essentially, fire was the discriminating agent in preventing white pine from competing on
the infertile, dry substrates of the Yellow Dog Plains. These studies have demonstrated
that there are inextricable links between geomorphology, soil texture (an indirect proxy
of soil moisture), and forest succession, which result in a given fire regime for a
particular landscape.

There is no shortage of evidence that fire was common before European
settlement in the sandy regions surrounding the Evart Upland. In an analysis of GLO

reports for Roscommon and Crawford Counties, Whitney (1986) found that ~ 7.3% of the

138

total area of sections surveyed made reference to burned patches, the majority of which
were in pine-dominated stands. In the Grayling outwash plains region, GLO surveyors
noted that 3% of the land area was recently burned, and several fires covered thousands
of acres (Albert, 1995). These findings suggest that fires were indeed frequent and
episodic on sandy substrates in the region surrounding, and within, the Evart Upland.
The intensity and spatial patterns of fire likely varied with respect to landform, but as
suggested by Whitney (1986), fire occurrence in the region was probably more related to
vegetation type than natural firebreaks. This notion is particularly important in the
context of this study because the Evart Upland is topographically very undulating, a trait
that is considered detrimental to the efficient spread of fire. However, soils in the upland
are also excessively drained and supported white pine-mixed hardwood forest prior to
European settlement, a combination that probably promoted episodic fire (Comer et al.,
1998) (see Fig. 2.15, study area chapter). As such, if fires were indeed episodic in the
study area (as the wide range of radiocarbon dates suggests), they may have also been
patchy and localized, since the highly undulating topography of the Evart Upland

probably constrained fires to small areas.

5.3 A Warmer and Dgyer Climate During the Mid-Holocene

It is well documented that climate has fluctuated significantly in the Great Lakes
region during the Holocene (the past 10,000 years) (Davis et al., 2000). Many
paleoenvironmental proxies have been used to demonstrate such variability, e. g., pollen
analysis (Delcourt et al., 2002), lake sediment cores (Dean, 1997), paleohydrologic

records (Booth and Jackson, 2003), and geomorphic studies of increased dune

139

mobilization (Arbogast and Packman, 2004) and fluvial incision (Arbogast et al., 2008),
to name a few. Of particular interest to this discussion are other studies that have
documented time intervals of dry climate or aeolian activity that closely match the two
radiocarbon dates obtained from buried soils in the Evart Upland.

In order to make reasonable comparisons between the radiocarbon dates in this
study and other climatic events cited in the literature, it must first be recognized that such
comparisons can only be made within a regional context. Booth et a1. (2006) found a
significant correspondence between paleohydrologic records from two bogs in the
Midwest, separated by ~1000 km: Minden bog in Salinac County, MI, and Hole bog in
Cass County, MN. As such, it is herein considered reasonable to reference proxy data of
Holocene climatic variability obtained within ~1000 km of the Evart Upland, but most
emphasis will be placed on those within 500 km of the upland. The following discussion
will highlight some examples of what others working in the region have found with
respect to Holocene climate, particularly during the time intervals that correspond to the
two radiocarbon dates found in this study. To reiterate, the two radiocarbon dates

obtained in this study yielded ages of 6840 i 30 and 920 d: 20 cal yrs ago.

W

In a study of lake sediments from Elk Lake, MN, Dean (1997) used
concentrations of silt-sized quartz and Na in lake cores as indicators of increased aeolian
activity. Since Elk Lake is a closed system (as are kettles), Dean attributed the
abundance of well-sorted, silt-sized quartz to aeolian deposition, presumably deflated and

deposited in Elk Lake during drier climatic intervals. Fluctuations in % Na content in the

140

 

lake sediments (the primary source of which is liberated from plagioclase feldspar via
hydrolysis) were used to track periods of former moisture availability (Fig. 5.1).
Therefore, drier periods are indicated in core samples by increased levels of Na relative
to the total detrital fraction, since less plagioclase is decomposed during periods of low
available moisture. Among the many cycles of decreased moisture availability during the
Holocene, a distinct dry interval is recorded by an increase in Na at ~6900 (varve years
before present) (Fig. 4.34). The results of Dean’s analysis (1997) confirm that the mid-
Holocene in Minnesota (8,000-5,500 years ago) was much winder and probably
drier/dustier than it is today (Dean, 1997).

A likelihood of drier and warmer climate during the mid-Holocene may also be
evident in the nearby landforms of the Muskegon River (Arbogast et al., 2008). Using
radiocarbon dates obtained from basal peats in paleomeanders of the Muskegon River,
Arbogast et a1. (2008) reconstructed the geometry and chronology of terraces in the
valley. Evidence from the sequence of terraces in the valley suggested that a significant
period of incision occurred in the Muskegon River sometime after 8,000 cal. years BP,
which the authors assumed was in response to a warmer and drier climate, and
subsequent decreases in discharge. Indeed, as new seismic profile and lake core evidence
from Lewis et a1. (2007) shows that Lakes Huron and Michigan experienced a
hydrologically closed lake low-stands at about this time (~7,900 years BP), presumably
driven largely driven by a warm, dry Holocene climate.

Further north, in the Baraga dune field in Baraga County, in Michigan’s western
upper peninsula, Arbogast and Packman (2004) obtained OSL (optically stimulated

luminescence) dates from vegetated dunes to reconstruct the history of their formation.

141

‘70 Na

0 0.2 0.4 0.6 0.8
0 ‘ L L l

 

 

OOmVQOIAOJM

d

 

Thousands of Varve Years (ka) Before 1980

 

 

11 'V‘vv'vv‘r—VI—T—IV‘tVTiV

0 0.5 1 1.5 2
Na-MAR

Figure 5.1- A Holocene record of aeolian activity from Elk Lake, MN. Black shading represents %
Na (higher values indicate increased dryness). Line pattern shading indicates mass accumulation
rate. The red bracket highlights a cycle of increased aeolian activity that corresponds with the oldest
radiocarbon date obtained in this study. Diagram after Dean (1997).

142

Their OSL dates indicated that aeolian activity in the Baraga dune field occurred in the
middle Holocene, from 7.0 to 6.0 ka (thousands of years before present). They cited two
possible mechanisms for the onset of dune formation: 1) increased westerly winds
associated with the 8.2 ka dry/cold climate event, as verified by evidence from lake
sediments in Elk Lake, MN (Dean et al., 2002), and/or 2) destabilization of surfaces due
to increased fire activity associated with Jack pine on sandy, flat, substrates.

In a similar study, Arbogast et a1 (2002) obtained samples for OSL dating from
dunes in the Newberry dune field, in the eastern upper peninsula of Michigan. The OSL
dates in their study indicated that aeolian activity occurred between 7.0 and 5.5 ka.
Fluctuating climatic conditions alone (i.e. warmer and dryer) were revoked as the sole
mechanism responsible for the onset of dune formation. Instead, they suggested that
isostatically driven regressions in the Lake Superior and Lake Michigan basins caused a
drastic drop in watertables in the region surrounding the Newberry dune field. Because
the dune field formed in a relatively thick sequence of sandy glacial sediments, they
suggested that even a small increase in temperature and dryness could combine with the
drop in groundwater to destabilize surface sediments and trigger the onset of aeolian
activity (Arbogast et al., 2002).

In a study comparing two sediment cores obtained from bogs in Michigan and
Minnesota, Booth et a1. (2006) used testate amoeba analysis to reconstruct the
paleohydrology at each site. Among other episodes of widespread drought indicated in
the analysis, their findings document the presence of a large drought interval between 900
and 1000 cal. years BP (Fig 4.35). They associated this drought interval with widespread

changes in vegetation and fire regime in southeastern Michigan.

143

Age(BP)
24002200200018001600140012001000 800 000 400 200 0

Widespread
.1
10 Hole Bog, drought

events

i 2.9

‘5‘

Depth to water table (cm)

 

 

 

 

 

Figure 5.2- Diagram showing the comparison of hydroclimate histories reconstructed from Hole Bog,
MN (Blue) and Minden Bog, MI (red) for the past 2500 yrs after (Booth et al., 2006). Both records
show a prominent dry interval around 1000 cal. years BP (highlighted yellow). This time period
corresponds to the radiocarbon date of 920 cal. years BP obtained in this research.

Other studies (Booth and Jackson, 2003; Davis et al., 2000; Dean et al., 2002)
also confirm periods of dry/warm climate and episodic drought throughout the Holocene
in the Great Lakes region; however, it is not the purpose of this discussion to provide a
comprehensive comparison to other published dates, because only two were obtained
during this research. The main point, in the context of this study, is that sandy upland
areas, such as the Evart Upland, would have been susceptible to climatic fluctuations
throughout the Holocene. Indeed, the radiocarbon dates obtained in this study (6840 i 30
and 920 i 20 years BP) overlap with periods of increased dryness during the Holocene.

As the highest and driest parts of the regional landscape, sandy uplands like the Evart

Upland are closer to t he threshold of disturbance from destabilizing agents such as fire

and wind than are adjoining regions/areas.

5.5 The Possible Role of Climate and Fire in Destabilizing Sandy Surfaces

There is unequivocal evidence in the literature suggesting that, historically, sandy
substrates in Michigan promoted vegetation types susceptible to fire (Brubaker, 1975;
Whitney, 1986; Leahy and Pregitzer, 2003; Arbogast and Packman, 2004). However, as
noted by Arbogast et a1. (2004), the role of fire as a destabilizing mechanism for surface
soils is contested. The root of this debate centers on the question of how quickly
stabilizing surface vegetation re-grows, and stabilizes the surface, after fire. The
efficiency of re-growth depends on the type of vegetation and general climate of the
region. For example, fire-adapted species may rapidly re-establish ground cover and
quickly stabilize soils after fire, but not if such a fire is associated with cold-climates
when rapid regrowth is not possible.

It has long been recognized that recently burned surfaces are temporarily
hydrophobic because of the addition of water repellent organic compounds released from
plant materials during fire (Savage, 1974; DeBano et al., 1979). Fire-induced soil
hydrophobicity increases the potential for runoff and soil erosion by water (Dragovich
and Morris, 2002; Gabet, 2003). Newer evidence suggests that the same compounds
responsible for fire-induced soil-hydrophobicity also increase the potential for wind
erosion (Ravi et al., 2006). Ravi et a1 (2006) carried out experiments on well-sorted,
clean sandy sediments in a wind tunnel that were artificially coated with a fatty acid

commonly found in plants (palmitic acid), which is released upon burning. Their results

145

indicated that fire-induced water repellency results in a decrease of threshold velocity for
wind transport, thereby enhancing the efficacy of wind erosion on sandy soil. In a wind
erosion study of semi-arid ponderosa pine forests of the western U.S., Whicker et al.
(2008) measured the amount of wind-eroded dust emanating from both burned and
thinned areas, as opposed to non-burned areas. Their results indicated that thinned areas
of the pine forest were just as susceptible to wind erosion as burned areas, and that both
emitted more wind-eroded dust than non-burned areas.

Although recent studies of the relationship between fire, wind-erosion, and dust
deflation are primarily focused on the arid and semi-arid portions of the U.S., it is
possible that similar processes were operating during the Holocene in the Great Lakes
region, especially on dry sandy upland areas, such as the Evart Upland, which would
have been more susceptible to the destabilizing mechanisms of drought, fire, and

subsequent wind-erosion.

5.6 Paleoenvironmental Siggificance - Summary

Fire was probably frequent on dry, sandy landforms, such as the Evart Upland,
throughout the Holocene because dry, sandy substrates allow for occasional fire, which
tends promote fire-prone species, and subsequent feedbacks. The presence of charcoal in
buried soils on the Evart Upland confirms that this landscape indeed has experienced fire
throughout the Holocene, and probably also throughout the Late Pleistocene. Indeed,
GLO reports of other sandy regions of Lower Michigan surrounding the Evart Upland
confirm that expansive areas of recently burned forest were common at the time of the

surveys (~1850’s) (Simard and Blank, 1982; Whitney, 1986). These reports confirm that

146

fire was a natural (and probably frequent) occurrence on sandy substrates in Lower
Michigan. Given the preponderance of sandy substrates and undulating topography in
the Evart Upland, fires may have been frequent, but localized and patchy.

In the context of the Evart Upland, linking the occurrence of fire to Holocene
climate proxies is simplistic. However, given that there are many lines of evidence
indicating that the Great Lakes region experienced significantly warmer and drier climate
during the Holocene, such a link is also realistic. Being the highest and driest parts of the
regional landscape, sandy uplands like the Evart Upland are close to the threshold of
disturbance from destabilizing agents such as fire. As such, even the smallest
fluctuations in Holocene climate would have impacted sandy, dry uplands more than
other sites in this part of Michigan.

As the data in this thesis suggest, silt deposition on the Evart Upland has occun'ed
episodically throughout the Holocene, and contemporaneously with fire. It is unclear
whether fire or a warm/dry Holocene climate was the actual causal mechanism triggering
destabilization of surfaces in and around the Evart Upland. In either case, deposits of
aeolian silt and charcoal in kettle bottoms on the Evart Upland record a history of

landscape evolution, greatly impacted by disturbance mechanisms such as wind and fire.

147

CHAPTER 6 - Conclusions

W

The Evart Upland, a ~68 km2 dome-like upland of rolling hills in north-central
Lower Michigan, contains an abundance of tOpographically closed basins (kettles).
These kettles likely formed as a complex arrangement of stagnant ice, leftover after the
Lake Michigan and Saginaw ice lobes retreated from the region, was surrounded and
variously buried by thick deposits of outwash. Subsequent meltout of the buried ice
resulted in a ubiquitous assemblage of kettles, many of them dry, set within a landscape
dominated by sand. These kettles have subsequently acted as accumulation basins,
preserving records of landscape stability and instability via the sediments that are
contained within their center-bottom portions.

The majority of kettles (53/59) sampled on the Evart Upland contain lenticularly-
shaped, silty deposits in the center-bottom portion (i.e., the deepest part) of the
depression. Silty deposits in kettle bottoms here are anomalous. Therefore, the purpose
of this research was to determine the most likely geomorphic origin(s) of these silty

deposits, and assess their paleoenvironmental importance.

6.2 Characteristics and Ori ° of Si] Kettle Bottom De osits

     

6.2.1 Aeolian Silt

Kettles in the Evart Upland have variously accumulated sediments since final
landscape stabilization. Indeed, nearly all of the kettles (53/59) investigated in the Evart

Upland contain small (< 1 % of the kettle area) silty deposits in the bottom center of the

148

depression; these sediments may extend to ~l m depth. Overall, these silty deposits are
texturally much finer than the surrounding backslope soils, which are sandy. On average,
kettle bottoms contain 44 % more clayfree silt than the surrounding backslopes.
Furthermore, the silt in kettle bottoms has a modal grain size of 18 um, whereas the silt in
backslopes has a modal grain size of 45 um. Finally, these silty deposits do not extend
up the surrounding backslopes, as would be expected if downslope sorting processes
were responsible for the silt. Because these silty deposits are anomalous with respect to
the surrounding backslope soils, as well as the preponderance of sand in the region, I
conclude that their origin is most likely aeolian.

Soils on three flat, geomorphically stable surfaces on the Evart Upland show an
increase in silt content in the upper (~0-50 cm) profile. These sites exhibit a 10% silt
increase in the upper profile, and in each case, fine silt (2-20 mm) increases more than
medium (20—35 mm) or coarse silt (35 um). Furthermore, fine silt is also the modal grain
size in kettle bottom deposit samples. I interpret this evidence to indicate that silt was
added to flat stable sites fi'om above, via aeolian deposition, and later, pedogenically
assimilated into the profile, this process diminishing with depth. Similar scenarios were
proposed for the preservation of aeolian silt in other flat landscape positions in northern
Lower Michigan (i.e. the Buckley flats (Schaetzl and Hook, 2008) and the Grayling
Fingers region (Schaetzl, 2008). Such an increase in silt in the upper profiles of flat,
stable sites offers supporting evidence that the silty kettle bottom deposits have an
aeolian origin, and confirms that aeolian processes have impacted soils in this landscape.
This evidence also suggests that the degree to which aeolian silt has been preserved and

incorporated into soils here may have been influenced by the slope (susceptibility to

149

runoff) or long-term stability of the surface. Surfaces with even the slightest slope may
not have preserved aeolian silt because it would have been easily washed downslope, and

spatially concentrated in the nearest basin.

6.2.2 Slopewash Sand

Kettle bottoms have also accumulated various amounts of sand within the
otherwise silty deposits; this sand presumably originated from the surrounding
backslopes during periods of increased slopewash deposition. Recent evidence (Fig 4.11)
and the depthplot diagrams (Figs. 4.20 - 4.35) indicate that kettle backslopes have been
subject to erosion at different times, and to different intensities, in the past. Each kettle
investigated with depth preserved a unique record of depositional events in the kettle,
ranging from intervals of pure silt (loess) deposition to sandy (slopewash) deposition
from the surrounding slopes, attesting to the variation in processes that were operative on

this landscape in the past.

6.2.3 Buried Soils

Various numbers of buried soils exist in each of several kettles investigated. The
soils are indicative of episodes of former landscape stability and instability. In some
kettle bottoms, a soil formed within a silty interval of sediment and was subsequently
buried by slopewash, whereas in other kettles, such a soil was buried again by loess. At
one site in particular (site 53), the lowermost buried soil (which formed in the underlying
sandy substrate sediment), was buried by loess (Figs. 4.35 and 4.37), indicating that a
period of stability pre-dated at least some of the aeolian silt depositional periods on the

Evart Upland. The variation in depositional sequence among kettles suggests that the

150

buried soils in kettle bottoms cannot be correlated among basins, and that some kettles

were experiencing significant erosion, while others were stable. Similarly, some kettles
were collecting loess while others were not, as indicated by the different amounts of silt
preserved within each kettle bottom deposit. As such, loess deposition in the study area

was probably both spatially and temporally discontinuous.

6.3 Silt Deposition Contemporaneous with Holocene Fires

The silt comprising the majority of kettle bottom deposits was probably deposited
contemporaneously with, or shortly after, fire events. Evidence for this includes
fragments of charcoal found within buried soils and throughout the profiles of silty kettle
bottom deposits. Samples of charcoal from two sites were radiocarbon dated. These two
dates span the middle to late Holocene (6840 i 30 and 920 i 20 cal. years BP).
Furthermore, these fires may have been intense, as indicated by the presence of scorched,
amorphous silica grains found in the kettle bottom deposits (Fig. 4.40). Such fires may
have liberated surficial sediments, thereby facilitating the aeolian redistribution of silt

into local kettle bottoms.

6.4 Kettle Bottom Silt is Local

The silt mineralogy of kettle bottom deposits was compared to the autochthonous
silt found in backslope soils. The similarity of the X-ray diffraction patterns between the
two types of sediments suggests that the silt deposited in kettle bottoms is locally

redistributed silt, perhaps winnowed out of surfaces disturbed by fire.

151

WM

It is well known that fire is a significant factor that regulates forest succession.
For example, some researchers in northern Lower Michigan have outlined significant
differences in forest succession pathways among different geomorphic landforms, which
are influenced largely by topography, soil type (an indirect proxy of soil moisture), and
ultimately, fire history (Host et al., 1987; Host and Pregitzer, 1992). An important
conclusion of their work is that landform types in Michigan control species composition
at the site level, as well as long-term disturbance patterns such as fire, the understanding
of which is traditionally obtained fi'om quantitative analysis of original land survey
records (Whitney, 1986; Leahy and Pregitzer, 2003).

Sandy upland areas, like the Evart Upland, would have been susceptible to
periods of drought, increased warmth, and fire during the Holocene. Consequently, the
Evart Upland has preserved long-term records of fire disturbance via the aeolain, silty
sediments that are contained in kettle bottoms. Such records may be useful in developing
an understanding of long-term disturbance histories in regions with sandy upland
landforms susceptible to fire, where traditionally, an understanding of such histories is
primarily restricted to analyses of GLO reports (i.e., short term records) (Leahy and
Pregitzer, 2003; Whitney, 1986). Finally, this research demonstrates that, on sandy
upland landforms of Lower Michigan, fire may have been a significant geomorphic agent

capable of liberating surficial sediments and facilitating silt deflation.

152

APPENDIX

153

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0N oN mm CN 6; h; 9mm wd cg 0.3m 5N
vN ON AN m; M; ed mom 9w m._ oNoN 0N
9v 9m v6 oN mg od 5mm mé— mg w.©NN mN
9m 5N On A; m; md Ohm N.: m._ m.mwN VN
mN _.N wN v4 _._ md 98 am _.— _.vom MN
o.m mN mm M: h; _.o N.ww NA: 5; m.©wN NN
QM 0m o6 mN vN v.0 fimw WE vN QSN _N
o.m N.v m6 v6. N.m md Q: Qw— N.m NumMN ON
N4» Wm Nam Wm m.m We WE. wd— Wm OSBN @—
mé Nam 9v _.N a; _.o wNw m.n_ a; 5mg 5
~.m ad Ym wN vN N.o Vow NN— vN w._MN 2
Wm 0.». 9v mN N.N _._ New 0.2 N.N 09% @—
m.m o.m vé vN wN od mew ~.m_ wN VNVN NE
EN oN 5N v; M." ed Woo N.w mg mNoN :
w.m m.m 06 m6 mm Nd ode m.wN 0.x @3— A:
9N vN QM wN 5m _.o Néw ~.N_ 5m vNoN a
w.m 9N wN m; o; _.o 0.3 vd— o._ waN w
N.N w; N.N .4 ad od Q; NS 0.0 QEN N.
5N 0N _.m _.N w; w._ «awn wd w; odwm c
_.m _.m v.0 Tm Qm _.0 men. 93 a6 VdVN m
9v o6 be mé 9v Yo wdb v.3 9v deN v
we N.w wd o.m mN _.o N65 m._N mN v.08 m
iv N6 w.w ON N.w To v.50 _.VN m.w :18 N
m6 9v v.0 0.0 No no 9% _.mN m6 mNmN _
E: om - mm 8: mm - N E: N - NE E: N— - N 8: N .. N6 8: OOONéoS E: oooN - on OMEN E: NV 8: 055am
Em D Em 2 Em m Em m> >20 U eaam D> uaam Em >30 mag—Z

 

 

 

 

 

 

 

 

 

 

 

 

.8? 239::— vofiumes 53:. n mag—’2 .0383 be .593; as 38398 9.2—E: Z .3383. one—manna he 53. .33er -—.< «San.

33.—Em one—8.3m -Sa: .53—

154

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Z 2 3. 2 N... mo 3.x «.2 an as: a
3“ am an an em 3 new 0.2 on Neg em
ca 3. E S. E 2 59 IN am 0me mm
3 i on 3 3. 3 E” as 3 £2 E
3” 3 am an 3. to as” an 3. an: mm
3 3 mm Z 3 3 3w 3 3 E: 9.
E E 2” M: 3 od Is as 3 E: E
Z 2 2 2 I o.o cg S E was a.
ma 3 an 2 3 do a; a: 2 $8 a.
3 Z S M: S o.o new we 5 22 t.
3 on we 3 3 to ME a: an 33 3
3 an 5 an 2 IV 3 n8 3 $2 a.
2 M3 3 3. 3. so new a: 5. Sam 3.
as 3 N4: 3 3 od 3% 3:. o.” 32 a.
Z 2 2. am 3 3 new h: 3 0me a.
3 Z .3. M: Z to Sm n: 2 use 2.
mm 3 an E 3 3 now as f 2% 8
3 3 3.. 2 9m 3 new 2: 3 non am
5. 3 no 2. 2. no was 3: S. 28 mm
X 3 an N... am to 5: <2 an ER 2
2V 3 3 E 3 3 NS mom 3 we: 8
we 3” 3V 3 3 9o 2» 3; S 32 mm
3 2 on 3 3 2 new 2: 3 3% X
3. Z 2. 3 Z to men 0.2 2 3% mm
2 3 Z I 3 ed 5% 3 1 38 NA
3 3 2 f 3 2 0.8 E 3 Sam 2
we 3 2 S S no new as S 2.2 om
3 mm 3 mm w.m o.o mom «.2 a... no: 8
2 3 3 S 2 to 3w 3 2 33. mm

am D am 2 am m E .9 50 u 25 u> saw :5 EU 23: 23am

 

 

 

 

 

 

 

 

 

 

 

 

325:8 -3. as;

 

155

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a.w WO_ WmN 0.2 O.: NO O.wN WOO v. : a.ma wN
Wm O.m W2 NNE a.m_ WO N._m WVM N4; v.mON bN
WN vN O.v vN ON WO Wow W: O.N OONM ON
v.5 We N.a ad ad EO NNO WbN O.m WE.— mN
We O.m WaE 5.2 5.2 _.O va Wow W: W: VN
fin _.w Wafi _.m_ N: _.O N.mm N.Om WE “new MN
v.5 WK. W3 a.O_ a.a 0.0 WDV WNV _.O_ W52 NN
We WK. N.a_ Wo— W: _.O a._m mom Omfi Nae EN
Wm mo WNN WmN va _.O a.w afim N.mm N.Nm ON
N6 N8 Wm— a.: QB N.O afim adv v.2 O.mO~ a—
ON WK. a.: W3 0.2 0.0 Odm adv Q: N.Nw w—
ON 5a N.mN N.a~ W: _.O TON W3 W5 N.wm S
We Wm WNN N.a~ a.ON 0.0 N.NN Wow WEN Wmm 2
ON 5w a._N ONE WE WO WON adm Wm: vaa N—
aO ON v.2 O.w WK. NO 2% Wmm Wu Wa: I
as Wa OéN O.: WE _.O VON me m: WE. O—
Ww Wa WON W2 W: _.O VON O.Nm W: Woa a
Wm Wm 5: ma WO_ a.O NNm W2” O.: O.Na~ w
Wm We ab— O.E v.3 ~.O WNv bNV Nav— bég A.
NO ON WNN Wafi VON N.O _.mN flow WON W: O
Wm WV WI Wa Wa N.O WOO NuaN Ya NOON m
O.m _.m Wu ON v.2 _.O use a.ON v. 2 Ram: v
_.m Wm WN_ 5a N.w WO amm WNm Wm vNa— m
ON N.a >.mN WE W2 _.O WVN O.am me— vgo N
56 56 a4: Wa Wm 0.0 O.¢m OHM O.a W2: E
Edom- mm 53mm- N 5: mN-N_ 5: N_ -N 5:N. N.O 5=OOON.OOO_ 5: OOON-Om 5=Om.N 5: NV 5: mEma—am
Em D Em E Em h Em m> N3:0 U 65am U> 15am Em .35 mg:

 

 

 

 

 

 

 

 

 

 

 

 

.33 232.3 cox—wt? 535 u mg: .9359. .3 2.3.83 3 63.833 52.5.5 .8358 3232. 53:5 233— ..8 5a.. :25on -N.< 935—.

835.3 3351 5325 2:3— -Ban Bay—

156

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Wm OO 0.0N WO~ 0.0_ _.O O.NO WOO Yv— WVO 5O
WO 0.0 WO— 0.0~ WE WO N.vO 0.0v OO— WO: OO
YO Wh 0.0_ WO— WO~ YO 0.00 WOv 0.0_ W_O_ OO
OS Wm O.NN WO— OO— _.O YVN WOO 0.0_ WOO O
_.O WO 92 O4; WN. 0.0 _.Ov O6“. O.N_ WOO OO
0.0 0.0 WO_ WO_ N.O_ 0.0 WNv WOv YO_ W; NO
YO Wm O.NN WOO 0.0_ NO WVN WOO YO_ WOO O
_.O Wm O.VN N.NN WON 0.0 W3 W3 0.0N W; OO
WO N.O 0.0N 0.0_ YE 0.0 W_O WOO WE W_O Ov
WO YO N.ON WE WO— N.O WhN N.OO WO_ 0.00 wv
YO WO O.w_ WO~ WE 0.0 0.00 W: 0.0_ WOO Sq
OS OO OO— YNO WO— 0.0 O._v ORV 0.0_ WNO~ Ov
YR. WO O._N W: Wm. 0.0 0.0N WOO WO— NNO Ov
WO Wm O.NN YO. WNN WO O._N WOO WON WK. 3
OO WO YNN 0.0~ 0.0_ 0.0 _.ON WOO NJ: N.ON. Ov
O8 OO N.O_ WOO WN_ _.O WOO OSV WN_ N.Ow Nv
0.0 YO 0.0_ O.: WO 0.0 0.0v 0.0v _.O_ N.OO 2“
YO YO WM: O.N_ WO- N.O 0.0v YOv 0.0_ WV: Ow
:. Wu 0.0_ WOO WE WO _.OO _._O WO— N.OO OO
0.0 0.0 WVN WON WON WO W: WOO WON 0.0v OO
0.0 OS 0.0N W: YO. N.O WOO 0.00 WO_ YOO 5O
WO NB _.O~ N.O_ 0.0_ _.O WVO WOV W: WOO OO
WK. Wm 0.0_ N.O_ W: _.O YOO O.~O W: WNN. OO
NS 0.0 0.0N _.ON 0.0_ 0.0 WO_ N.OO 0.0N WOV ”O
0.0 0.0 O.: WO_ YO_ WO OSO 0.0v WO_ WO: OO
0.0_ WOO O.NN WN_ _.O NO WvO 0.00 WO _.OO NO
0.0 NO 0.0N N.ON 0.0N _.O 0.0N OOO ~._N WOO _O
Wm Wm YO_ WN_ WO_ WO 0.0v wi O.: WON_ OO
N.O OS N.O_ W3 O.N_ WO 0.0¢ an O.N_ WhO_ ON
Em D Em E Em m Em m> >20 0 Onam U> cuam Em has mg: o-QEam

 

 

 

 

 

 

 

 

 

 

 

 

125653 -~.< ~35.

 

157

BIBLIOGRAPHY

158

Albert, D. A. 1995. Regional landscape ecosystems of Michigan, Minnesota, and
Wisconsin: a working map and classification. St. Paul, Minn.: US. Dept. of
Agriculture, Forest Service, North Central Forest Experiment Station.

Almendinger, J. E. & S. Hobbie (1992) Influence of terrestrial vegetation on sediment-
fonning processes in kettle lakes of west-central Minnesota. Quaternary
Research, 38, 103-116.

Arriaga, F. J ., B. Lowery & M. D. Mays (2006) A fast method for determining soil
particle size distribution using a laser instrument. Soil Science, 171, 663-674.

Barrett, L. (2001) A strand plain soil development sequence in Northern Michigan, USA.
Catena, 44, 163-186.

Beierle, B. D., S. F. Lamoureux, J. M. H. Cockbum & I. Spooner (2002) A new method
for visualizing sediment particle size distributions. Journal of Paleolimnology, 27,
279-283.

Bennett, M. & N. F. Glasser. 1996. Glacial geology: ice sheets and landforms.
Chichester; New York: Wiley.

Beuselinck, L., G. Govers, J. Poesen, G. Degraer & L. Froyen (1998) Grain-size analysis
by laser diffractometry: comparison with the sieve-pipette method. Catena, 32,
193-208.

Booth, R. K. & S. T. Jackson (2003) A high-resolution record of late-Holocene moisture
variability from a Michigan raised bog, USA. Holocene, 13, 863-876.

Booth, R. K., M. Notaro, J. E. Kutzbach & S. T. Jackson (2006) Widespread drought
episodes in the western Great Lakes region during the past 2000 years:
Geographic extent and potential mechanisms. Earth and Planetary Science
Letters, 242, 415—427.

Buurman, P., T. Pape & C. C. Muggler (1997) Laser grain-size determination in soil
genetic studies 1. Practical problems. Soil Science, 162, 211-218.

Chen, Y., M. Song & M. Dong (2002) Soil properties along a hillslope modified by wind
erosion in the Ordos Plateau (semi-arid China). Geoderma, 106, 331-340.

Davis, M., C. Douglas, R. Calcote, K. L. Cole, M. G. Winkler & R. F1akne(2000)

Holocene climate in the western Great Lakes National Parks and Lakeshores:
Implications for future climate change. Conservation Biology, 14, 968-983.

159

Dean, W. E., R. M. Forester & J. P. Bradbury (2002) Early Holocene change in
atmospheric circulation in the Northern great plains: An upstream view of the 8.2
ka cold event. Quaternary Science Reviews, 21, 1763-1775.

DeBano, L. F., R. M. Rice & C. E. Conrad (1979) Soil heating in Chaparral fires: effects
on soil properties, plant nutrients, erosion, and runoff. USDA Forest Service
Research Paper, PSW-145.

Folsom, M. M. 1971. Glacial geomorphology of the Hastings quadrangle, Michigan. 5, v,
166 leaves. Michigan State University. Dept. of Geography, 1971.

Ham, N. R. & J. W. Attig (1996) Ice wastage and landscape evolution along the southern
margin of the Laurentide Ice Sheet, north-central Wisconsin. Boreas, 25, 171-186.

Host, G. E. & K. S. Pregitzer (1992) Geomorphic influences on ground-flora and
overstory composition in upland forests of northwestern lower Michigan.
Canadian Journal of Forest Research, 22, 1547-1555.

Host, G. E., K. S. Pregitzer, C. W. Ramm, J. B. Hart & D. T. Cleland (1987) Landform-
mediated differences in successional pathways among upland forest ecosystems in
northwestern Lower Michigan. Forest Science, 33, 445-457.

Jenny, H. 1941. Factors of soil formation; a system of quantitative pedology. New York,
London: McGraw-Hill.

Johnson, M. D., D. M. Mickelson, L. Clayton & J. W. Attig (1995) Composition and
genesis of glacial hummocks, western Wisconsin, USA. Boreas, 24, 97-116.

Kleiss, H. J. (1970) Hillslope sedimentation and soil formation in northeastern Iowa. Soil
Sci. Soc. Am. Proc., 34, 287-290.

Konert, M. & J. Vandenberghe (1997) Comparison of laser grain size analysis with
pipette and sieve analysis: a solution for the underestimation of the clay fraction.
Sedimentology, 44, 523-535.

Kufmann, C. (2003) Soil types and eolian dust in high-mountainous karst of the Northern

Calcareous Alps (Zugspitzplatt, Wetterstein Mountains, Germany). Catena, 53,
211-227.

Lamoureux, S. (2000) Five centuries of interannual sediment yield and rainfall-induced
erosion in the Canadian High Arctic recorded in lacustrine varves. Water

Resources Research, 36, 309-318.

Leahy, M. J. & K. S. Pregitzer (2003) A comparison of presettlement and present-day
forests in northeastern lower Michigan. American Midland Naturalist, 149, 71 -89.

160

Lewis, C. F. M., C. W. Heil Jr, J. B. Hubeny, J. W. King, T. C. Moore Jr & D. K. Rea
(2007) The Stanley unconformity in Lake Huron basin: Evidence for a climate-
driven closed lowstand about 7900 p14 sC BP, with similar implications for the
Chippewa lowstand in Lake Michigan basin. Journal of Paleolimnology, 37, 43 5-
452.

Loizeau, J. L., D. Arbouille, S. Santiago & J. P. Vernet (1994) Evaluation of a wide range
laser diffraction grain size analyser for use with sediments. Sedimentology, 41,
353-361.

Malo, D. D., B. K. Worcester, D. K. Cassel & e. al. (1974) Soil-landscape relationships in
a closed drainage system. Soil Sci. Soc. Am. J., 38, 813-818.

Michigan Department of Information Technology, Center For Geographic Information.
Online Content. Available online at http://www.michigan.gov/cg accessed
February 26, 2007

 

Mikesell, L., R. J. Schaetzl & M. A. Velbel (2004) Homblende etching and
quartz/feldspar ratios as weathering and soil development indicators in some
Michigan soils. Quaternary Research, 62, 162-171.

Milne, G. (1936) Normal erosion as a factor in soil profile development. Nature, 138,
548.

NCAA, Data from NOAA Midwestern Regional Climate Center. Illinois State Water
Survey at the University of Illinois at Urbana-Champaign
under the Institute of Natural Resource Sustainability. Available online at
http://mrcc.isws.illinois.edu. accessed 11/6/08

NRCS, Plant fact sheet: Sweetfem, Comptonia peregrina. United States Department of
Agriculture, Natural Resources Conservation Service. Available online at
http://p1ant-materials.nrcs.usda.gov/ accessed 12/2/2008

Paton, T. R., G. S. Humphreys & P. B. Mitchell. 1995. Soils: a new global view. New
Haven: Yale University Press.

Pennock, D. J. & E. De Jong (1987) The influence of slope curvature on soil erosion and
deposition in hummocky terrain. Soil Science, 144, 209-217.

Rieck, R. L. 1976. The glacial geomorphology of an interlobate area in southeast
Michigan: relationships between landforms, sediments, and bedrock. [4], xci,

[218] leaves. Michigan State University. Dept. of Geography, 1976.

--- (1979) Ice stagnation and paleodrainage in and near an interlobate area. Papers Mich.
Acad. Sci, Arts, Letters, 11, 219-235.

161

Rieck, R. L. & H. A. Winters (1993) Drift volume in the Southern Peninsula of Michigan
- a prodigious Pleistocene endowment. Physical Geography, 14, 478-493.

Savage, S. M. (1974) Mechanism of Fire-Induced Water Repellency in Soil. Soil Sci. Soc.
Am. J, 38, 652-657.

Schaetzl, R. J. (1996) Spodosol-Alfisol intergrades: bisequel soils in NE Michigan, USA.
Geoderma, 74, 23-47.

--- (2008) The distribution of silty soils in the Grayling Fingers region of Michigan:
Evidence for loess deposition onto frozen ground. Geomorphology: in press.

Schaetzl, R. J. & S. L. Forman (2008) OSL ages on glaciofluvial sediment in northern
Lower Michigan constrain expansion of the Laurentide ice sheet. Quaternary
Research: in press.

Schaetzl, R. J. & J. Hook (2008) Characterizing the silty mantle on the Buckley Flats
outwash plain: Evidence for loess in NW Lower Michigan. Physical Geography:
in press.

Schaetzl, R. J ., B. D. Knapp & S. A. Isard (2005) Modeling soil temperatures and the
mesic-frigid boundary in the central Great Lakes region, 1951-2000. Soil Science
Society of America Journal, 69, 2033-2040.

Schaetzl, R. J. & B. N. Weisenbom (2004) The Grayling Fingers region of Michigan:
Soils, sedimentology, stratigraphy and geomorphic development.
Geomorphology, 61, 251 -274.

Simard, A. J. & R. W. Blank (1982) Fire history of a Michigan jack pine forest. Mich.
Academ, 15, 59-71.

Soil Survey Staff, Natural Resources Conservation Service, United States Department of
Agriculture. Soil Survey Geographic (SSURGO) Database for Osceola County,
MI. Available online at http://soi1datamart.nrcs.usda.gov accessed 02/17/2007

Sperazza, M., J. N. Moore & M. S. Hendrix (2004) High-resolution particle size analysis
of naturally occurring very fine-grained sediment through laser diffractometry.
Journal of Sedimentary Research, 74, 736-743.

Walker, P. H. & R. V. Ruhe (1968) Hillslope models and soil formation. 11. Closed
systems. Trans. 9th Intl Congr. Soil Sci, 58, 561-568.

Whitney, G. G. (1986) Relation of Michigan's presettlement pine forests to substrate and
disturbance history. Ecology, 67, 1548-1559.

162

—1

Zhang, G., J. Germaine, R. Martin & A. Whittle (2003) A simple-sample mounting
method for random powder x-ray diffraction. clays and clay minerals, 51, 218-
225.

163