Embed Size (px)
Citation preview
Article Type: Original Article
Title: Phytolith Concentration and Morphotypes in Modern Soils of the Columbia Basin,
USA as Indicators of Vegetation Composition and Cover
Authors: Reyerson, P.E.a, 1, corresponding author
Affiliation Address:
aDepartment of Geography
St. Cloud State University
Stewart Hall 344
720 4th Ave South
St. Cloud, MN 56301
USA
Present Address:
1Department of Geography
University of Wisconsin-Madison
160 Science Hall
550 North Park Street
Madison, WI 53706
USA
Phone: 320-260-0624
Email: [email protected]
Research interests: Quaternary paleoecology, paleoenvironments,
vegetation dynamics, climate change
Blinnikov, M.S.a
11
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Affiliation Address:
aDepartment of Geography
St. Cloud State University
Stewart Hall 344
720 4th Ave South
St. Cloud, MN 56301
USA
Phone: 320-308-2263
Email: [email protected]
Research interests: Quaternary paleoecology, phytolith analysis
Busacca, A.J.b
Affiliation Address:
bDepartment of Crop and Soil Sciences
Washington State University
245 Johnson Hall
P.O. Box 646420
Pullman, WA 99164
USA
Phone: 208-885-7505
Email: [email protected]
Research interests: Pedology, geomorphology, paleoclimatology
Gaylord, D.R.c
Affiliation Address:
22
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
cSchool of Earth and Environmental Sciences
Washington State University
1146 Webster Physical Sciences Building
Pullman, WA 99164
USA
Phone: 509-315-8127
Email: [email protected]
Research interests: Quaternary eolian-climatic interactions; clastic
sedimentary environments
Rupp, R.b
Affiliation Address:
bDepartment of Crop and Soil Sciences
Washington State University
405 Johnson Hall
Pullman, WA 99164
USA
Phone: 509-335-2381
Email: [email protected]
Research interests: remote sensing, GIS
Sweeney, M.R.d
Affiliation Address:
dEarth Sciences Department
University of South Dakota
33
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
414 East Clark Street
Vermillion, SD 57069
USA
Phone: 605-677-6142
Email: [email protected]
Research interests: Geomorphology, sedimentology, paleoclimatology
Zender, C.S.e
Affiliation Address:
eDepartment of Earth System Science
University of California-Irvine
3200 Croul Hall
Irvine, CA 92697
USA
Phone: 949-824-2987
Email: [email protected]
Research interests: Climate, surface-atmosphere interaction
ABSTRACT
The goals of this paper are to assess feasibility of using 1) total opal concentration (TOC)
of phytoliths in modern soils and 2) morphotypes in soil assemblages to infer
paleovegetation cover and composition. We sampled from 37 plots in undisturbed
grassland sites with different values of shrub and grass cover in the Columbia Basin,
southeastern Washington State, USA. Phytolith concentration in modern soils was
44
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
measured and all morphotypes counted. Estimates of vegetation cover were obtained
using in-field visual inspection and NDVI values derived from georeferenced
multispectral digital imagery. Vegetation composition by species was determined in-
field. Linear regressions were used to predict vegetation cover based on TOC in soils.
TOC in modern soils ranged from 0.99% to 4.28% of dry weight. In-field inspection and
NDVI values revealed direct positive relations between vegetation cover and TOC (0.21
R2 for in-field estimation and 0.49 R2 for NDVI). CCA was used to assess the degree of
similarity between field plots. The first two CCA axes explain about 32% of the total
variance in the data (28% for the 1st axis); the morphotypes-species correlations for the
axes are moderately high (Pearson correlation coefficient r=0.863 for the first axis, and
r=0.660 for the second axis), suggesting that phytoliths reflect vegetation plot
compositions with moderately high certainty. We conclude that phytoliths accurately
reflect modern vegetation in grasslands and shrublands on the Columbia Basin. Estimates
of vegetation cover and composition (from in-field and NDVI data) reflect annual and
inter-annual trends, while the phytolith assemblages from modern soils reflect the
vegetation averaged over a decadal to centennial scale. While the presence or absence of
an individual species cannot be detected with certainty from the phytolith assemblages, a
few regionally dominant genera of grasses can be identified using phytoliths.
Keywords
Grasslands, modern analogs, phytoliths, morphotypes, Columbia Basin, Palouse
INTRODUCTION
55
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
Modern assemblages of pollen, diatoms, macrofossils and opal phytoliths are frequently
used to guide paleoenvironmental reconstructions (e.g., Overpeck et al., 1986; Guiot et
al., 1993; Fredlund and Tieszen, 1997b; Whitmore et al., 2005). Since the pioneering
work of Fredlund and Tieszen in the Great Plains (1994, 1997a), phytolith assemblages
have proven to be a powerful method by which to infer paleovegetation composition in
arid environments where pollen is not well preserved. More recently, Dehlon et al.
(2003), working in the middle Rhone valley of France, employed phytolith analogs to
infer changes in Holocene Mediterranean vegetation. Bremond et al. (2005) demonstrated
the utility of using phytoliths to reconstruct vegetation along a forest-savanna transect in
southeastern Cameroon in Africa. Lu et al. (2005) provided compelling evidence that
phytolith morphotypes can be used to reliably estimate mean annual precipitation,
temperature and other climate variables on regional scales in China.
Blinnikov et al. (2002) used the modern analog method to reconstruct
paleovegetation of the interior Pacific Northwest (Columbia Basin) from an
approximately 100,000 yr record of phytoliths preserved in the Palouse loess. However,
the modern dataset used in that study did not provide adequate quantitative estimates of
the vegetation cover. Further, Blinnikov et al. (2002) chose to distinguish forest,
grassland and shrubland vegetation of the large region, not specific grassland types. The
study presented in this paper addresses a gap in knowledge regarding phytolith
production rates within a mixed grassland and shrubland on the modern Columbia Basin.
By focusing on a relatively limited area of study we believe we can more confidently
resolve local-scale differences in vegetation cover and composition than has been done
previously.
66
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
In this paper, we refer to ‘vegetation composition’ as the presence and abundance
of certain silica-producing plant species and ‘vegetation cover’ as the percentage of plant
canopy cover, which is estimated by in-field visual inspection and the normalized
difference vegetation index (NDVI). Many studies have demonstrated that vegetation
composition influences, at the very least, the presence and abundance of specific
phytolith morphotypes (Verma and Rust, 1969; Blinnikov et al., 2002; Reyerson, P.E.).
However, few studies have focused on the relation between vegetation cover and total
opal concentration (TOC) in modern soils (expressed as a percentage of dry weight of
extracted opal from the dry bulk soil). This relation between vegetation cover and TOC
remains unclear. A quantification of this relation can be addressed by comparing the
vegetation cover of a given area to its TOC. Ultimately, this will lead to more accurate
paleoenvironmental reconstructions.
The Columbia Basin is located in an arid portion of the U.S. Pacific NW behind
the rain shadow of the Cascade Mountains. Quaternary sedimentary deposits in the basin
have been the subject of many paleoenvironmental reconstructions where pollen,
phytoliths, faunal burrows and stable isotopes have been used as climate proxies
(Barnosky, 1985; Minckley and Whitlock, 2000; Whitlock et al., 2000; O’Geen and
Busacca, 2001; Blinnikov et al., 2001; Blinnikov et al., 2002; Reyerson, P.E.; Blinnikov,
2005; Stevenson et al., 2005). Because of high oxidation rates and generally low rates of
burial, few sedimentary repositories have existed on the Columbia Basin during the Late
Pleistocene to Holocene that have favored either pollen or macrofossil preservation.
Lakes are scarce, and the majority are less than 10 000 years old (Mack et al., 1976).
Minckley and Whitlock (2000) presented a regional modern pollen assemblage dataset
77
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
from which they assessed the spatial variability of pollen influx across the Pacific
Northwest; but their data came primarily from sites closer to the Cascade Mountains, and
not from the lower, drier Columbia Basin. In addition to this geographic limitation, their
study does not provide a record of potentially revealing phytolith populations that could
accompany and complement the pollen record.
In the absence of a pollen record, opal phytoliths have emerged as a powerful
proxy for determining vegetation composition in arid regions in North America (Fredlund
and Tieszen, 1994; Fearn, 1998; Kelly et al., 1998; Kerns, 2001; Blinnikov et al., 2001,
2002; Blinnikov, 2005). Many taxa (particularly grasses, forbs, shrubs, and trees) produce
opal phytoliths in their cell walls and intercellular spaces. Phytolith production is
taxonomically specific and genetically controlled (Hodson et al., 2005). The abundance
and persistence of phytoliths in soil are considered by many researchers to be directly
related to the plant biomass of silica-producing species and the amount of time the
vegetation existed at the site (Bartoli and Wilding, 1980; Fredlund and Tieszen, 1994;
Alexandre et al., 1997b; Blinnikov, 2005; Blecker et al., 2006). Phytoliths are presumed
to accumulate mostly in situ (Piperno, 1988, 2006), but remobilization of these particles
by surface sedimentary processes is a real possibility that must be considered during
analysis (Fredlund and Tieszen, 1994). In temperate, semi-arid areas such as the Great
Plains or the Columbia Basin, phytoliths can persist for thousands of years in soils under
a variety of pH and moisture conditions (Kelly et al., 1998; Blinnikov et al., 2001;
Blinnikov et al., 2002; Reyerson, P.E.). Phytoliths have been utilized to identify the
expansion of grasslands in the Eocene and early Miocene (Strömberg, 2004). While
Strömberg (2004) was not able to use modern analogs as an aid in reconstruction, it is
88
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
prudent to calibrate estimations of the paleoecological make-up of the area in question
when possible for greatest accuracy. Modern phytolith assemblages and distributions for
a given region must be recognized before paleoenvironmental reconstructions can be
undertaken with any certainty, however (Bowdery, 1998; Carnelli et al., 2001; Lu and
Liu, 2003; Blinnikov, 2005).
This study investigates the relations between the modern vegetation from selected
grass- and shrubland sites in the Columbia Basin and phytolith assemblages. The goal of
this comparative study is to provide a calibration tool for on-going and future Quaternary
paleoclimatic research in this basin. Correlations between the volume of plant
biomass/cover and soil opal on modern plots and documentation of the relations between
plant species and soil morphotypes from those same plots will permit researchers to
reconstruct both the composition and structure of past vegetation. Such reconstructions
can be used to establish vegetation boundary conditions for input into numerical
paleoclimate models. Such paleobotanical boundary conditions are necessary to estimate
dust flux rates from the paired (sand: silt/clay) eolian system since the last glacial
maximum (LGM) (Mahowald et al., 2006), when episodes of heightened eolian activity,
including loess generation and deposition, have been recorded (Sweeney et al., 2005).
Two specific research questions are explored in this paper: (i) if assemblages of phytolith
morphotypes from modern soils collected on standard vegetation plots (6 x 6 m) within a
relatively small geographical area (ca. 30,000 km2) reliably reflect modern grassland and
shrubland vegetation composition; and (ii) if measures of TOC in modern topsoil be used
to calculate estimated plant percent cover obtained through remote sensing (e.g., NDVI)
99
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
on the same plots. The results from this study will be used to constrain the accuracy of
paleobotanical data we will provide to numerical climate modelers.
Study Area
The Columbia Basin is a structural and topographic low and subprovince of the
Columbia Intermontaine Physiographic Province (Freeman et al., 1945; Baker et al.,
1991). It is a basin surrounded by the Blue Mountains on the south, the Cascade
Mountains on the west, the Okanogan Highlands on the north, and the Rocky Mountains
on the east (Fig. 1). The Columbia Basin lies south of the maximum extent of the
Cordilleran Ice Sheet, and largely south and east of the Columbia River. The total area of
the basin exceeds 160 000 km2.
Large portions of the Columbia Basin in SE Washington State are blanketed by
thick layers of Quaternary loess (windblown silt). Known informally as the Palouse
Loess, these deposits have been molded into rolling hills that dominate the topography of
much of eastern Washington (Baker et al., 1991; Busacca and McDonald, 1994). The
source of the loess is primarily slackwater sediment from glacial outburst floods (Bretz,
1969; Baker et al., 1991; Sweeney et al., 2005) that accumulated primarily in lowland
areas within the Columbia Basin throughout the Quaternary. These slackwater
sedimentary deposits largely occur along the Columbia River and are upwind and
southwest of the Palouse. Other, secondary source areas that could have contributed
eolian silt include the Late Miocene-Pliocene Ringold Formation, Pleistocene outwash
sediment, and Quaternary alluvium from the Columbia River (Busacca and McDonald,
1994). Modern soils that developed on the loess are mostly Mollisols or Aridisols (Baker
et al., 1991). These surface soils form along a strong climatic (primarily moisture)
1010
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
gradient that correlates with a decrease in precipitation across the Columbia Basin from
NNW to SSE (Bolig et al., 1998). A detailed chronology of the loess deposition is well
established (Baker et al., 1987; McDonald et al., 1988; Baker et al., 1991; Busacca et al.,
1992; McDonald and Busacca, 1992; Busacca and McDonald, 1994; O’Geen and
Busacca, 2001; Sweeney et al., 2004; 2005). The use of luminescence techniques and
tephrochronology have greatly facilitated reconstruction of the loess depositional history
(Berger, 1991; Busacca et al., 1992; McDonald and Busacca, 1992; Berger et al., 1995;
Richardson et al., 1997; Richardson et al., 1999; King et al., 2001).
METHODS
Field data collection
A total of 37 modern soil samples were collected for the study. Sample sites were chosen
primarily on the basis of their undisturbed nature. Very few such natural sites exist in this
heavily agricultural region, which greatly limited our choices. We eventually chose four
nature preserves along a roughly south to north climatic gradient (temperature and
precipitation), all of which have substantial areas covered by pre-agricultural vegetation.
The preserves are aligned with relatively cooler and wetter preserves in the north and
warmer, drier preserves to the south. Two field sites were located in southeastern
Washington (Marcellus and Kahlotus Ridge preserves of the Washington Department of
Natural Resources), and two in northeastern Oregon (Lindsay Prairie and Boardman
Preserves of the Nature Conservancy). Table 2 provides exact geographic coordinates.
1111
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
Despite our best efforts to utilize pristine sites, all preserves contain some
invasive species, most notably annual cheat brome (Bromus tectorum L.). However, due
its relatively recent introduction, B. tectorum is expected to have a minimal impact on the
modern phytolith record. Our reasons for this assumption are twofold. First, soil phytolith
assemblages are the products of gradual accumulation over many growing seasons that
may encompass hundreds of years (Fredlund and Tieszen, 1994). Second, while the mean
residence time for phytolith persistence in the soil (Fredlund and Tieszen, 1994) is not
known for our study areas, recent studies have suggested that the rate of phytolith
dissolution may be very high in temperate grasslands (Alexandre et al., 1997; Derry et al.,
2005; Blecker et al., 2006). Therefore, sizeable accumulations of B. tectorum phytoliths
should take many growing seasons.
Each soil sample was obtained as an aggregate of 10 random pinches within a 6x6
m square plot. Each plot was larger than 4 m2 plots used in Daubenmire (1970), so that
they could be located from the aerial surveys conducted as part of this study. While the
plots were clustered geographically into 4 nature preserves, we assumed each plot to be
an independent sample. To achieve this, we attempted to maximize the vegetation
variability within each preserve to include as many combinations of grassland and
shrubland types and cover as possible in our plots. Each square plot was aligned to the
magnetic north with a Brunton compass.
The UTM coordinates of the southwest corner of each study plot were recorded
using a professional mapping grade handheld Trimble GPS receiver with differential
correction to the nearest permanent station, which allowed us to achieve about 1 m
horizontal positional accuracy. Next, a permanent spike with a tag was placed on the
1212
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
southwest corner of the plot. The linear boundaries were then entered into a GIS polygon
layer.
Overhead color images of each study plot were obtained using a hand-held
camera mounted on a 12-ft. tall pole to create a permanent archive of plant cover and
composition. Plant cover and composition and percent plant cover for each species were
estimated in the field at the peak of the flowering season (mid-May 2003). Due to budget
constraints, airborne surveys of the same plots were flown only in May, 2004 using a
digital multispectral camera with 10cm on-the-ground resolution. All evident plant
species that had been identified and recorded in May, 2003 were deemed appropriate
given the similarity in growing seasons between the two years of sampling. Each plant
was assigned to one of 4 functional groups (shrub, grass, forb or legume). The region of
study has very few C4 grasses and none were encountered on field plots; thus, all the
grasses in this study are C3.
Ten random pinches were taken from each study plot, placed in a plastic bag and
thoroughly mixed. These samples were collected from plant detritus-free surfaces to a
depth of less than 2 cm. Any evident plant detritus was manually removed before
sampling the soil. The total amount of sample procured from each plot was
approximately 50 g. No attempt was made to subsample within each plot.
Phytolith processing
Phytoliths were extracted using the modified wet oxidation technique of Pearsall (2000).
Soils were left in a drying oven at approximately 80 ºC overnight and then weighed and
1313
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
placed in 100 ml beakers. Carbonates and most organics from the soil were removed by
digesting with 10% HCl for 15 minutes and then boiling in 70% HNO3 with a pinch of
KClO3 for 1.5 hours. After deflocculation in 5% (NaPO3)6 solution, the samples were
floated in a heavy liquid solution of ZnI2 calibrated to a specific gravity of 2.3 g cm-3.
This step was repeated at least three times to ensure that most of the opal in the sample
was retrieved. The ratio between the resulting dry weight of retrieved opal and the
original dry soil weight yielded the total opal concentration (TOC) and is expressed as a
percentage in Table 2.
Visual inspection of the resulting opal under a light microscope verified that the
extraction was reasonably pure (less than 5% of observed fragments were not biogenic in
origin). Steps were taken to recover as much opal from the samples as possible, and to
avoid any loss. Microscopic analysis of the decanted liquid in a few samples did not
reveal loss of phytolith material.
Individual opal phytolith shapes (morphotypes) were counted systematically using
a Leica optical microscope, model DMLB, at 400x magnification mounted in immersion
oil type A, which permitted the rotation and movement of the phytoliths on the
microscope slide to augment the estimation of true three-dimensional shapes. Twenty-six
distinct morphotypes were identified and recorded for each sample (Fig. 2). Digital
images and permanent slides were created for future reference. Morphotypes were
scanned in a row-like fashion, from left to right and from top to bottom of each slide. A
total of 300 grains were counted for each sample. Any object whose size was between 20
µm and 50 µm was counted, and was assumed to be biogenic silica in origin, unless it
was apparent that it was not (e.g., dark organic material or distinctly shaped thin mica
1414
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
plates). Morphotype classes were modified from Blinnikov (2005). At the conclusion of
the counting, seven of the originally defined morphotypes were removed from
consideration because they occurred only in trace amounts. These morphotype classes
were not considered statistically significant because they occurred in such low numbers
as to have little impact. For example, only 7 saddle (SA) morphotype grains were counted
for the entire study, out of a total of over 11 000. Further, these morphotypes were not
considered primarily diagnostic; that is, their presence or absence did not indicate the
presence or absence of a key species. See Table 1 for final morphotypes, nomenclature
and species associations.
Image processing
Additional estimates of long-term vegetation cover/biomass were obtained by measuring
Normalized Difference Vegetation Index (NDVI) on bands 4 and 3 of color-infrared
digital aerial imagery, using a standard algorithm (Jensen, 1996, p. 182). A chartered
plane was flown in May 2004 to obtain GPS-referenced digital images with 4 channels
(blue, green, red, and near-infrared) comparable in spectral sensitivity to the first four
bands of Landsat ETM+. Image resolution was approximately 0.16 m2 on the ground per
pixel. GPS-obtained vector files of the sample plot boundaries were matched to the
existing digital orthoquads of the four nature preserve areas to extract 37 grids of pixels
corresponding to the 37 plots. NDVI values were estimated for each plot by averaging
pixel values (from about 250 pixels per plot) to yield a mean NDVI value for each plot.
1515
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
Statistical Methods
Two statistical methods were used to compare phytolith assemblages to recorded
vegetation cover and composition. First, linear regression was utilized to determine if
there is a relation between total opal concentration (TOC) in soils and vegetation cover.
Second, Canonical Correspondence Analysis (CCA) (ter Braak, 1986; McCune et al.,
2002; run in PC-ORD) was utilized to evaluate the overall relations between the plots
(rows) and morphotypes (columns) in the main matrix and plots (rows) and percent cover
by plant taxa (columns) in the environmental matrix. We chose 10 morphotypes for this
analysis including straight rectangular plates, two types of rondels (pyramidal and keeled
or horned), short wavy trapezoids, Stipa-type bilobates, grass seed epidermis long cells,
regular long cells (rods), silicified hairs and trichomes, epidermal non-grass phytoliths,
and blocky forms. This level of phytolith classification was sufficiently broad to allow
unambiguous identification of all of these forms. More detailed classifications could be
used, but would result in much longer time necessary to count a statistically valid number
of grains. The plant taxa used were Agropyron spicatum, Bromus tectorum, Festuca
idahoensis, Poa secunda, Stipa spp., Asteraceae forbs, and shrubs. Together, these
accounted for about 95% of all plant cover, with the remainder mostly in representatives
of Liliaceae and Fabaceae families, which do not produce phytoliths.
CCA was chosen because it allows simultaneous ordination of morphotypes
(species) and plots with plant taxa (environmental variable) and plots in a single analysis.
This method also allows quick visualization of relations between morphotype and plant
taxa data in a single graph. We did not attempt to use CCA to investigate actual
1616
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
environmental (climate or soil) variables in this study given the limited coverage of
suitable pristine sample sites in the Columbia Basin.
RESULTS
Total Opal Concentration in Soils Estimates of Modern Vegetation Cover and NDVI
Mean annual precipitation decreases from NNW to SSE, and mean annual temperature
increases from north to south across the study area. As a result, Lindsay Prairie and
Boardman preserves are drier and warmer sites, while Marcellus and Kahlotus Ridge are
cooler and wetter (Fig. 1). The northernmost Marcellus preserve north of Ritzville, WA is
located within the Artemisia tridentata–Festuca idahoensis association of Daubenmire
(1970, 1972) and is cooler and wetter than the Kahlotus Ridge site. The Marcellus site
also has the highest percentage of shrubs and Festuca cover. The Kahlotus Ridge
preserve north of Kahlotus, WA is located in the middle of a grassland immediately north
of the Snake River in Agropyron–Festuca association of Daubenmire (1970, 1972) and
has very few shrubs. Boardman preserve, OR is a large grassland area south of the
Columbia River, and it is more arid than the northern preserves. It contains high
percentages of Stipa and Agropyron bunchgrasses and some Artemisia shrubs on canyon
slopes. The soils are more sand-enriched compared to Marcellus and Kahlotus Ridge.
Lindsay Prairie, OR is a small preserve south of Boardman, similar to the Boardman
preserve and the Kahlotus preserve in that both contain sandy soils and relatively high
proportions of Stipa and Agropyron grasses and some shrubby plots. Lindsay Prairie has
1717
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
some unique communities present such as Oryzopsis communities not found in the other
preserves examined in this study. Lindsay Prairie has a similar climate Boardman
Preserve.
As stated above, each plot evaluated in this study was considered separately. In
this way, each plot could be compared to other plots, regardless of proximity. Each
sample plot was selected for its own distinctive vegetation attributes in order to maximize
overall variability of vegetation cover and composition.
Modern vegetation cover values on the study plots range from 9% to 98% (Table
2). Grasses are the most dominant functional group, averaging 34% of composition by
cover. Of the four preserves, Kahlotus Ridge reported the highest overall vegetation
cover (52%), the highest proportion of grasses (44%) and the lowest proportion of shrubs
(4%). Marcellus reported a mean vegetation cover of 51%. Its proportion of grasses to
shrubs was more balanced than at Kahlotus Ridge, with 29% covered by grasses and 19%
by shrubs. Lindsay Prairie averaged 48% vegetation cover, with 42% grasses and 5%
shrubs present. One plot at Lindsay Prairie exhibited the overall highest vegetation cover
of any plot in the study, at 98% (L4). This is relevant because it shows the high degree of
intra-preserve variability, suggesting a low interrelatedness between the plots. Boardman
Preserve’s average vegetation cover (32%) was the lowest for any plot; plot BO7 from
that preserve also exhibited the lowest percent vegetation cover of all plots (9%). The
average proportion of grasses and shrubs from Boardman Preserve was 23% and 8%,
respectively. As was expected, as mean annual precipitation (MAP) increased and mean
annual temperature (MAT) decreased, the percent of vegetation cover also increased.
1818
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
The density of vegetation of each plot is reflected in the total opal concentration
(TOC). TOC for all 37 modern samples averaged about 2.5% (Fig. 6). Overall, Boardman
Preserve and Lindsay Prairie had the lowest TOC values (1.4% and 2.2% respectively).
Marcellus yielded the highest average TOC at 3.3%, while Kahlotus Ridge reported
3.0%. The highest TOC was observed at M10 (4.3%; Marcellus preserve) and lowest at
BO9 and BO5 (0.99%; Boardman preserve). All TOC values >3% corresponded to plots
from either Kahlotus Ridge or Marcellus preserve (the northern preserves), which are
characterized by higher vegetation (and grass) cover. Plot L4 (Lindsay Prairie) - with the
highest vegetation cover reported - had only 2.3% TOC. However, this plot is composed
mostly of shrubs, not grasses. Shrubs, overall, tend to produce a lower proportion of
phytoliths when compared to grasses. The five plots from which we identified the lowest
TOC values were all located in the Boardman preserve. These five plots possessed higher
average proportions of shrubs and/or weedy annual grass species (Bromus tectorum,
Vulfia octoflora), which likely would have contributed less opal to the soils than native
grasses.
A linear regression analysis was used to determine the relation between
vegetation cover on plots (as established visually in the field by an experienced observer)
and TOC. The best fit least squared linear regression equation was VC = 10.7(TOC) +
20.1, where TOC is total opal concentration, and VC is vegetation cover (Fig. 3). As can
be seen in Figure 3, there is a direct positive relation albeit a weak one (R2 = 0.21),
between TOC and vegetation cover. Regressions determined for grasses or shrubs alone
yielded lower R2 values (0.14 for grasses only and 0.07 for shrubs). Despite our best
efforts, the ranges of shrub cover values were rather inconsistent. Most plots yielded
1919
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
shrub cover values of <10% and only 5 had values >25%. No plots contained between 12
and 25% of shrub cover. More continuous ranges of cover were obtained for grass cover
with TOC values falling between 8% and 68%, with no major gaps.
In general, in-field estimates of plant cover may be misleading, because they are
often based on qualitative visual observations and, in our case, only a single visit to each
plot in May of 2003. Remote sensing analyses, in contrast, can provide an objective
measure of photosynthetically-active biomass, a factor not detectable by the human eye
(Fig. 4). Values of vegetation biomass expressed as a normalized difference vegetation
index (NDVI) were measured from digital aerial images. Utilizing a separate linear
regression, the same TOC values from modern soils were implemented to predict NDVI
values. The resulting graph is presented in Fig. 5. The correlation between NDVI values
and TOC was more strongly positive than we determined using the in-field inspection
method. The regression equation is NDVI = 0.0475(TOC) – 0.432 (R2 = 0.49, p<0.001).
The robustness of the strong correlation between TOC and NDVI should be assessed by
applying our methods or equivalent techniques in other regions. This predictive relation
between TOC and NDVI can help reconstruct paleo-NDVI from TOC measurements.
NDVI can also be used to constrain the vegetation boundary conditions in dust
entrainment models (e.g., Zender et al., 2003).
Analysis of Modern Morphotypes
The majority of the modern samples possessed similar morphotype assemblages (Fig. 6).
Most morphotypes were found on every plot, but often in different proportions. Most
2020
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
assemblages were dominated by plates (PR; 10% to 26%), short trapeziforms (WS; 8% to
25%), long cell rods (LR; 5% to 24%), oblong/oval rondels (RO; 4% to 15%), and keeled
rondels (RK; 2% to 11%), which are all grass phytoliths. The presence of morphotypes
for non-grasses were likewise similar between the plots, with a greater proportion of
polygonal forms (EP; 12% to 29%) than blocky forms (BL; 3% to 12%). (Refer to Table
3 for morphotype nomenclature.) The ubiquitous presence of morphotypes on most plots
was expected, considering the relatively homogenous species composition on all study
plots in the study area. All plots were located in non-forested, upland sites with the same
seven regionally dominant species of grasses and shrubs (Agropyron spicatum, Festuca
idahoensis, Poa sandbergii, Stipa comata, Artemisia tridentata, and two Chrysothamnus
spp., Table 2).
Grass phytoliths, which include thirteen morphotypes identified in this study,
made up between 62% and 84% of all phytoliths in soil samples (Fig. 6). The PR type
accounts for 10% to 26% of all morphotypes, and is the dominant short cell. However,
the PR type is common in many grasses and is not diagnostic of any specific plant species
or genus. The WS (short crenate) type ranges from 8% to 25% of all morphotypes. The
third most common morphotype in the assemblages was the LR (smooth rod) type, which
ranged from 5% to 23% of the total morphotype population. The phytoliths of microhairs
(HH) and trichomes (HT) ranged from 0% to 4%. Non-grass types, made up of EP
(polygonal) and BL (blocky) forms, range from 24% to 27% of all phytoliths.
The eigenvalues for the first CCA axes were λ1=0.027 and λ2=0.004. The species-
environment correlations for the axes are moderately high (Pearson correlation
coefficient r=0.863 for the first, and r=0.660 for the second). In the CCA environmental
2121
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
matrix, the first axis appears to be most positively correlated with the Bromus tectorum
and Stipa cover (the former is an annual weed, the latter is a native grass associated with
dry soils). The same axis is negatively correlated with the cover of native mesic grass
species Poa, Agropyron and Festuca. Therefore, the axis represents a moisture gradient.
The second axis was most positively correlated with forbs; Agropyron and Festuca cover
and most negatively correlated with Poa, Stipa and shrub cover. Thus, the second axis
represents a grassland to shrub steppe gradient. Essentially, both axes capture one
complex environmental gradient of shrubland-grassland transition across the region based
on moisture. A tight clustering appears for the samples on the CCA biplot showing two
first axes of the main matrix (Fig. 7). The samples from the warmer and drier preserves
(Lindsay Prairie and Boardman) cluster in the lower right of the biplot, while the samples
from cooler and wetter preserves (Kahlotus Ridge and Marcellus) tend to be grouped
more to the upper left. Data from the northern preserves revealed a higher abundance of
cooler/wetter climate species, such as Festuca, while data from the southern preserves
revealed higher proportions of warm/dry climate adapted Agropyron, Artemisia, and
Stipa occurring at overall lower densities.
Additionally, CCA helped us assess the relative position of each plot with respect
to different morphotypes from the main CCA matrix. Wavy trapezoids, bilobate Stipa-
type and silicified grass hairs, hair bases and trichomes tend to occur on plots from more
southern preserves with a higher incidence of Stipa, Oryzopsis (a related genus to Stipa)
and Bromus. The oval and keeled rondels and plates, on the other hand, are better
represented in the northern preserves on plots with a high incidence of either Festuca,
Agropyron, or both. Epidermal non-grass phytoliths occur on plots with higher
2222
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
percentages of shrubs and forbs. Blocky forms occur on plots where Festuca and shrubs
are common, including a few from Marcellus preserve. Overall, the CCA results agree
well with the expected pattern of phytolith production from modern plants (Table 1;
Blinnikov, 2005).
DISCUSSION
Total Opal Concentration in Soils Estimates of Modern Vegetation Cover and NDVI
Total opal concentration (TOC), expressed as percent of extracted plant opal relative to
the total dry weight of soil, is positively correlated to the observed vegetation cover on
the plots (as percent ground cover by species and plant functional group). TOC is also
positively correlated to Normalized Difference Vegetation Index (NDVI) as measured by
color-infrared aerial imagery. Although many studies use phytolith assemblages to
reconstruct the composition of vegetation found at a particular site (Bartolome et al.,
1986; Fisher et al., 1995; Blecker et al., 1997; Fredlund and Tieszen, 1997a; Dehlon et
al., 2003; Blinnikov, 2005), few have focused on the TOC in the soil (Prebble, 2003b;
Blecker et al., 2006). With accurate TOC values, it is possible to reconstruct
paleovegetation cover, which can lead to a clearer understanding of vegetation ecotone
dynamics and climate. Indeed, there is precedence for determining TOC values in
paleoreconstruction studies. As demonstrated by previous investigations, TOC is a
reliable indicator of paleovegetation, especially when grassland species are the dominant
vegetation, but temperate deciduous and coniferous forests were also applicable (Verma
2323
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
and Rust, 1969; Wilding and Drees, 1971). These two papers revealed that TOC values
from grassland prairie sites in Minnesota and Ohio, contained 3 to 10 times, respectively,
the TOC values of forest top soils. Paleovegetation cover in the Columbia Basin is an
important factor to measure because it so strongly influences eolian activity and soil
formation. Vegetation cover also may directly reflect soil moisture (Morrow and Friedl,
1998). Several studies have linked sand dune reactivation and loess deposition with
drought (Muhs et al., 2000; Mason et al., 2004; Sweeney et al., 2005; Miao et al., 2007,
Sweeney et al., 2007). Others have linked increased eolian deflation to reduced
vegetation cover (Snelder and Bryan, 1995; Arens et al., 2001; Mason et al., 2001;
Engelstaedter, 2003; Sweeney et al., 2007) and decreased root mass (Dong et al., 2001;
Wang et al., 2003). Increased surface roughness owing to a higher percentage of
vegetation cover also can increase drag on the wind, and protect soils from deflation
(Mason et al., 1999; Dong et al., 2003; Crawly and Nickling, 2003). Studies have shown
that increased soil moisture and vegetation cover leads to increased rates of pedogenesis
(Huang et al., 2001; Wolfe et al., 2001; Gustavson and Holliday, 1995).
Regression analysis revealed a significant positive trend. As TOC values
increased, the percent of vegetation cover also rose (Fig. 3). Thus, our TOC values helped
us predict the modern vegetation cover on our sample plots. Although we treated all plots
as independent samples, the plots in southern preserves (Lindsay Prairie and Boardman),
that evolved under warmer and drier conditions had lower overall vegetation densities
and also lower TOC values than their counterparts to the north. Thus, we do observe the
regional signal in the TOC values.
2424
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
When the TOC value for a given plot is over-predicted (such as BO2) or under-
predicted (such as L4), the disparity may be the result of short-term fluctuation in
biomass at the site. For example, analysis at plot L4 revealed a vegetation cover of 98%,
yet its TOC value was only 2.3%. TOC values reflect an averaging of vegetation densities
temporally. Therefore, the regression equation predicted the vegetation cover on L4 at
45%. However, 40% of the vegetation cover on L4 is explained by exotic annual Bromus
tectorum. While B. tectorum produces phytoliths, its contribution to the assemblage per
unit of covered area is likely to be much lower than for native bunchgrasses (e.g., A.
spicatum or F. idahoensis), given its much smaller biomass and tendency towards
interannual fluctuations in growth. If B. tectorum cover is ignored, the predicted value
(45%) is closer to that observed (58%). However, the full ramifications of ignoring
invasive species in a plot have not been thoroughly researched, and should be approached
with caution. Plot BO2, with 13% observed cover, yielded a TOC value of 2.1%, which is
roughly the median value for the entire dataset. This overabundance of TOC may suggest
that the observed plant cover is lower than suggested by phytoliths. Therefore, there may
be some inheritance of phytoliths on the plots from earlier growing episodes.
NDVI linear regression yielded a stronger correlation with TOC than that based
on visual observation of plants in the field. In both cases (NDVI and visual inspection)
measurements were performed within a single day (no repetitive estimates of cover or
NDVI). NDVI captures the difference in photosynthetic rates of plants as opposed to
recording plants covering a certain percentage of the ground surface. Higher
photosynthetic activity results in more active growth of plant tissues and should translate
into overall higher silica deposition in plants as well. Two sources of uncertainty play a
2525
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
role in estimating NDVI in our study: imprecise georeferencing of the aerial photos and
seasonality. The first is easier to estimate, because we were able to match the aerial
imagery with our rectified in-field square plot boundaries in a GIS. Since the
georeferencing positional accuracy is around 1 m, an offset by 1 m linearly in both X and
Y directions would allow 11 m2 out of the total 36 m2 of the plot be excluded from the
pixel count, and an additional 11 m2 to be brought in from the surrounding plot
vegetation. All of our plots, however, were chosen inside a larger patch of similar,
homogenous vegetation, so this effect was probably not very pronounced. The second
source of error is the fact that only one digital fly-over image was available per plot taken
on a specific day at the height of vegetative season for all preserves. We do not have
multidate or multiyear samples. However, the reasonably high linear relationship between
TOC and NDVI suggests that this was not a significant problem.
Modern Morphotypes Distinguish Shrublands from Grasslands and Dry Grasslands
from Mesic Grasslands
The overall diversity of phytolith morphotypes in our samples was relatively low. We
deliberately chose to use only a few broad categories of phytoliths in this study. While
more detailed morphotype classifications can be tested, our goal was to assess the broad
applicability of the method to resolve significant differences in plant cover and
composition. Most samples were dominated by the same seven morphotypes: plates (RP),
short trapeziforms (WS), keeled and oval rondels (RK and RO, respectively) and long
cell rods (LR) from grasses and blocky (BL) and epidermal polygonal (EP) forms from
2626
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
shrubs and forbs (Fig. 6). This is not surprising, given the relatively few dominant species
that are phytolith producers in the region (Blinnikov, 2005). No panicoid or chloridoid
species of Poaceae are dominant in the region. Thus, essentially all phytoliths in our
records originate from Fectucoid native (Agropyron, Festuca, Poa) and non-native
(Bromus tectorum) grasses. The bilobate forms of Stipa are not “true bilobates”, but
rather trapezoids of transitional type (Mulholland, 1989). A handful of true bilobates and
saddles were identified in the record and most likely were also contributed by Stipa
and/or Oryzopsis.
This study focused on a narrow environmental gradient ranging from dry
sagebrush communities to moderately mesic grasslands within a relatively small
geographic area. The much higher variability of morphotypes documented in the Great
Plains (Fredlund and Tieszen, 1994) and across China (Lu et al., 2005) is consistent with
the continent-spanning gradients in those studies. However, we did find important
differences in our assemblages highlighted in the CCA study.
CCA revealed that there is a good agreement between the vegetation growing on
the plots and morphotypes found in soil (Fig. 7). For example, plots with Stipa-type
bilobate phytoliths in soils are also the ones that support Stipa today. Plots with a high
incidence of silicified hairs (HT) corresponded with a high concentration of Bromus
tectorum, which has heavily silicified hairs. Plots with a high incidence of epidermal
polygonal phytoliths from shrubs are the ones with high shrub content today. Contrary to
our expectations, there was less correspondence between shrub cover and blocky forms.
Blinnikov (2005) reported that blocky forms occur primarily in Artemisia and perhaps in
other shrubs from Asteraceae family in the region. However, ‘blocky’ is a catch-all
2727
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
category for phytoliths that are large in size and 3-dimensional in appearance. Certain
types of grass bulliform cells also may resemble these. Keystone-shaped bulliforms,
which are very distinct, were not found in this study, but some blocky forms may have
been contributed by grasses. Also, our collection of forbs from the area is far from
complete. It is possible that some yet unstudied taxa contribute blocky forms in non-
shrub communities.
The most encouraging result from our research was that plants above ground seem
to contribute morphotypes in situ. Few studies exist that prove this connection (Kerns et
al., 2001). Generally, phytoliths are expected to provide a rather local signal, since it is
assumed that there is little post-depositional translocation (Piperno, 1988). Fredlund and
Tieszen (1994), however, found that in the Great Plains, only about 50% of phytoliths
from modern soils could be considered truly local (i.e., from within a few meters of a 4x4
m plot); the rest had been transported in from the surrounding one hectare. We
acknowledge that transport may be a legitimate problem, however our dataset indicates it
was not a major issue in our study. It seems that there is a strong correspondence between
the presence of certain phytolith producers with distinct shapes (e.g., Stipa) in our plots
that do deposit these primarily locally (within 6x6 m plots). In most cases, we did not
find a large number of foreign phytoliths that we could not link to the present vegetation.
Another issue with phytolith records in soils is inheritance. There were some plots
in our study where the morphotypes in the soils did not match well with the species
compositions on the plots For example, plot BO8 (Boardman preserve) yielded a high
incidence of blocky phytoliths presumably from shrubs, but no shrubs in the surface
record. However, history of the site (TNC staff, pers. comm.) indicates that a recent fire
2828
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
may have destroyed the pre-existing sagebrush on this site, which takes longer to recover
than grasses. Phytoliths in the upper 2 cm of the soil profile likely accumulated in the
past few decades, and thus more accurately reflect the longer-term period of vegetation
on the site, and not necessarily only current vegetation. This longer-term phytolith
accumulation can cloud the modern-day phytolith record in soils somewhat, but
acceptable for paleoecological studies where temporal resolutions are much larger
(Blinnikov et al., 2002).
We also found that annual exotic weeds (Bromus tectorum) may influence
modern-day phytolith records. Seed phytoliths, presumably from Bromus, and silicified
hairs common in this species were found on plots on which B. tectorum occurs with
regularity. This highlights the detective ability of phytolith analysis, but also dilutes the
assemblage of morphotypes from native grasses. Although due care was taken to sample
only native grasslands, the reality is that B. tectorum in the Columbia Basin ecosystem is
an ever-increasing problem, even on the preserves where our study was conducted.
However, the presence of B. tectorum doesn’t invalidate our study: this species has not
been present for a great enough time period to have a significant impact on the phytolith
record we examined.
CONCLUSIONS
The first attempt to quantitatively analyze phytolith soil assemblages and total opal
concentrations (TOC) in the Palouse area of the Columbia Basin, Washington State has
shown several benefits and limitations. The use of TOC as an indicator of vegetation
2929
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
cover and NDVI has been successful, but reveal only a weak positive correlation with the
vegetation cover and moderately strong correlation with the NDVI (R2 of 0.22 for plant
cover and 0.49 for NDVI). Morphotypes in soils indicate the above-ground plant taxa on
our 6 x 6 m plots. Our results show that TOC is important and morphotype assemblages
do reasonably approximate the vegetation on a given plot over a timescale of at least a
few decades. Short-term fluctuations in vegetation attributes are muted (averaged) in the
phytolith assemblages. Even though exotic weeds (Bromus tectorum) have been detected
in the modern assemblages it does not mitigate our results. Most assemblages from
grasslands and shrublands in our area are broadly similar. There are no morphotypes that
would distinguish one vegetation type from another, but overall, existence of even small
amounts (3-5%) of Stipa-type bilobates is a strong indicator that Stipa and Oryzopsis are
present in the modern vegetative cover. Likewise, the occurrence of epidermal polygonal
forms strongly suggests that shrubs are represented as well. Silicified hairs found in the
assemblages suggest the presence of Bromus tectorum, an exotic weed. More mesic
grasslands can be distinguished from more arid ones on the basis of higher proportions of
plates and blocky forms in the former, while the arid ones will have a higher incidence of
short trapeziforms and rods. Very few phytoliths were contributed by Asteraceae and
other forbs, although contribution to the assemblages by these species seems likely. The
issues of inheritance and translocation of phytoliths in post-depositional period were not
conclusively resolved in this study. Some of our morphotypes may have come from
outside the plots or were inherited from previous times (e.g., when shrubs were removed
in a crown fire), but we did not see significant evidence for this possibility. The issues
should be studied further in a series of controlled experiments with long-term mixtures.
3030
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
REFERENCES
Alexandre, A., Meunier, J., Lezine, A., Vincens, A., Schwartz, D., 1997. Phytoliths:
indicators of grassland dynamics during the late Holocene in intertropical Africa.
Palaeogeography, Palaeoclimatology, Palaeoecology 136, 213-229.
Arens, S.M., Baas, A.C.W., Van Boxel, J.H., Kalkman, C., 2001. Influence of reed stem
density on foredune development. Earth Surface Processes and Landforms 26,
1161-1176.
Baker, V.R., Greeley, R., Komar, P.D., Swanson, D.A., Waitt, R.B., Jr., 1987. Columbia
and Snake River plains, in: Graf, W.L. (Ed.), Geomorphic Systems of North America.
The Geological Society of America, Boulder, CO., pp. 403-468.
Baker, V.R., Bjornstad, B.N., Busacca, A.J., Fecht, K.R., Kiver, E.P., Moody, U.L.,
Rigby, J.G., Stradling, D.F., Tallman, A.M., 1991. Quaternary geology of the Columbia
Plateau, in: Morrison, R.B., (Ed.), Quaternary nonglacial geology; Conterminous U.S.
Geological Society of America, The Geology of North America, v. K-2, Boulder, CO.,
pp. 215-250.
Barnosky, C.W., 1985. Late Quaternary vegetation in the southwestern Columbia Basin,
Washington. Quaternary Research 23, 109-122.
3131
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
Bartoli, F., and Wilding, L.P., 1980. Dissolution of biogenic opal as a function of its
physical and chemical properties. Soil Science Society of America 44, 873-878.
Bartoli, F., 1983. The biogeochemical cycle of silicon in two temperate forest
ecosystems, in: Hallberg, R., (Ed.), Environmental Biogeochemistry. Proceedings of
the 5th International Symposium, Stockholm, 1981. Ecological Bulletins, Stockholm,
pp. 469-476.
Bartolome J.W., Klukkert, S.E., Barry, W.J., 1986. Opal phytoliths as evidence for
displacement of native Californian grassland. Madrono 33, 217-222.
Berger, G.W., 1991. The use of glass for dating volcanic ash by thermoluminescence.
Journal of Geophysical Research 96, 19705-19720.
Berger, G.W., Busacca, A.J., 1995. Thermo-luminescence dating of late Pleistocene loess
and tephra from eastern Washington and southern Oregon, and implications for the
eruptive history of Mount St Helens. Journal of Geophysical Research 100,
22361-22374.
Blecker S.W., Yonker, C.M., Olson, C.G., Kelly, E.F., 1997. Paleopedologic and
geomorphic evidence for Holocene climate variation, Shortgrass Steppe, Colorado,
USA. Geoderma 76, 113-130.
Blecker, S.W., McCulley, R.L., Chadwick, O.A. Kelly, E.F., 2006. Biologic cycling of
silica across a grassland bioclimosequence. Global Biogeochemical Cycles 20,
GB3023, doi: 10.1029/2006GB002690.
3232
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
Blinnikov, M.S., Busacca, A.J., Whitlock, C., 2001. A new 100,000-year record from the
Columbia Basin, Washington, USA., in (Meunier, J.D., Colin, F., Eds.), Phytoliths –
Applications in Earth Science and Human History. A.A. Balkema, Rotterdam, pp.
27-55.
Blinnikov, M.S., Busacca, A.J., Whitlock, C., 2002. Reconstruction of the late
Pleistocene grassland of the Columbia Basin, Washington, USA, based on phytolith
records in loess. Palaeogeography, Palaeoclimatology, Palaeoecology 177, 77-101.
Blinnikov, M.S., 2005. Phytoliths in plants and soils of the interior Pacific Northwest,
USA. Review of Palaeobotany and Palynology 135, 71-98.
Bolig, M., Frazier, B., Busacca, A., 1998. General soil map of Washington. Washington
State University, Pullman, WA.
Bowdery, D., 1998. Phytolith analysis applied to Pleistocene–Holocene archaeological
sites in the Australian arid zone. British Archaeological Reports International Series no.
695. Hadrian Books, Oxford.
Bremond, L., Alexandre, A., Véla, E., Guiot, J., 2004. Advantages and disadvantages of
phytolith analysis for the reconstruction of Mediterranean vegetation: an assessment
based on modern phytolith, pollen and botanical data (Luberon, France). Review of
Paleobotany & Palynology 129, 213-228.
3333
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
Bremond, L., Alexandre, A., Hély, C. Guiot, J., 2005. A phytolith index as a proxy of tree
cover density in tropical areas: Calibration with Leaf Area Index along a forest-savanna
transect in southeastern Cameroon. Global and Planetary Change 45, 277-293.
Bretz, J.H., 1969. The Lake Missoula floods and the Channeled Scabland. Journal of
Geology 77, 505-543.
Busacca, A.J., Nelstead, K., McDonald, E.V., Purser, M., 1992. Loess in the Channeled
Scablands of Washington State: tephra correlations. Quaternary Research 37, 281-303.
Busacca, A.J., McDonald, E.V., 1994. Regional sedimentation of late Quaternary loess
on the Columbia Plateau: sediment source areas and loess distribution patterns.
Washington Division of Geology and Earth Resources Bulletin 80, 181-190.
Carnelli, A.L., Madella, M., Theurillats, J.P., 2001. Biogenic silica production in selected
alpine plant species and plant communities. Annals of Botany 87, 425-434.
Crawley, D.M. Nickling, W.G., 2003. Drag partition for regularly-arrayed rough
surfaces. Boundary-Layer Meteorology 107, 445-468.
Daubenmire, R., 1970. Steppe Vegetation of Washington. Washington Agricultural
Experiment Station, Washington State University, Pullman, WA.
Daubenmire, R., 1972. Annual cycles of soil moisture and temperature as related to grass
development in the steppe of eastern Washington. Ecology 53, 419-424.
3434
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
Delhon C., Alexandre, A., Berger, J., Thiebault, S.T., Brochier, J.E., Meunier. J., 2003.
Phytolith assemblages as a promising tool for reconstructing Mediterranean Holocene
vegetation. Quaternary Research 59, 48-60.
Derry, L.A., Kurtz, A.C., Ziegler, K, Chadwick., O.A., 2005. Biological control of
terrestrial silica cycling and export fluxes to watersheds. Nature 433, 728-731.
Dong, Z., Gao, S., Fryrear, D.W., 2001. Drag coefficients, roughness length and zero-
plane displacement height as disturbed by artificial standing vegetation. Journal of Arid
Environments, 49, 485-505.
Dong, Z., Liu, X., Wang, H., Wang, X., 2003. Aeolian sand transport: a wind tunnel
model. Sedimentary Geology 161, 71-83.
Engelstaedter, S., Kohfeld, K.E., Tegen, I., Harrison, S.P., 2003. Controls of dust
emissions by vegetation and topographic depressions: An evaluation using dust storm
frequency data. Geophysical Research Letters 30, 271-274.
Fearn, M.L., 1998. Phytoliths in sediment as indicators of grass pollen source. Review of
Palaeobotany and Palynology, 103, 75-81.
Fisher, R.F., Bourn, C N., Fisher, W.F., 1995. Opal phytoliths as an indicator of the
floristics of prehistoric grasslands. Geoderma, 68, 243-255.
Fredlund, G.G., Tieszen, L.T., 1994. Modern phytolith assemblages from the North
American Great Plains. Journal of Biogeography, 21, 321-335.
3535
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
Fredlund, G.G., Tieszen, L.T., 1997a. Phytolith and carbon isotope evidence for late
quaternary vegetation and climate change in the southern Black hills, South Dakota.
Quaternary Research 47, 206-217.
Fredlund, G.G., Tieszen, L.T., 1997b. Calibrating grass phytolith assemblages in climatic
terms: application to late Pleistocene assemblages from Kansas and Nebraska.
Palaeogeography, Palaeoclimatology, Palaeoecology 136, 199-211.
Freeman, O.W., Forrester, J.D., Lupher, R.L., 1945. Physiographic divisions of the
Columbia Intermontaine Province. Annals of the Association of American Geographers
35, 53-75.
Gaines, E., Kagan, J., Hak, J.C. Oregon historical land cover. Retrieved December 1,
2004, from web site Gap Analysis Program, FTP download: ftp://ftp//gap.uidaho.edu/.
Guiot, J., de Beaulieu, J.L., Cheddadi, R., David, F., Ponel, P., Reille, M., 1993. The
climate in Western Europe during the last Glacial/Interglacial cycle derived from pollen
and insect remains. Palaeogeography, Palaeoclimatology, Palaeoecology 103, 73-93.
Gustavson, T.C., Holliday, V.T., 1999. Eolian sedimentation and soil development on a
semiarid to subhumid grassland, tertiary Ogallala and Quaternary Blackwater Draw
Formations, Texas and New Mexico High Plains. Journal of Sedimentary Research 69,
622-634.
Hodson, M.J., White, P.J., Mead, A., Broadley, M.R. (2005) Phylogenetic variation in the
silicon composition of plants. Annals of Botany 96, 1027-1046.
3636
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
Huang, M., Shao, M., Li, Y., 2001. Comparison of a modified statistical-dynamic water
balance model with the numerical model waves and field measurements. Agricultural
Water Management 48, 21-35.
Kelly, E.F., Blecker, S.W., Yonker, C.M., Olson, C.G., Wohl, E.E., Todd, L.C., 1998.
Stable isotope composition of soil organic matter and phytoliths as paleoenvironmental
indicators. Geoderma 82, 59-81.
Kerns, B., 2001. Diagnostic phytoliths for a ponderosa pine–bunchgrass community near
Flagstaff, Arizona. The Southwestern Naturalist 46, 282-294.
Kerns, B.K., Moore, M.M., Hart, S.C., 2001. Estimating forest-grassland dynamics using
soil phytolith assemblages and δ13C of soil organic matter. Ecoscience 8, 478-488.
King, M., Busacca, A.J., Foit Jr., F.F., Kemp, R.A., 2001. Identification of disseminated
Trego Hot Springs tephra in the Palouse, Washington State. Quaternary Research 56,
165-169.
Lu, H., Liu, K., 2003. Phytoliths of common grasses in the coastal environments of
southeastern USA. Estuarine, Coastal and Shelf Science 58, 587-600.
Lu, H., Wu, N., Yang, X., Jiang, H., Liu, K., Liu, T., 2005. Phytoliths as quantitative
indicators for the reconstruction of past environmental conditions in china I: Phytolith-
based transfer functions. Quaternary Science Reviews 25, 945-959.
3737
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
Mack, R.N., Bryant, V.M., Jr., Fryxell, R., 1976. Pollen sequence from the Columbia
Basin, WA: reappraisal of postglacial vegetation. American Midland Naturalist 95,
390-397.
Mahowald, N.M., Muhs, D., Levis, S., Rasch, P., Yoshioka, M., Zender, C.S., Luo, C.,
2006. Change in atmospheric mineral aerosols in response to climate: last glacial
period, preindustrial, modern, and doubled carbon dioxide climates. Journal of
Geophysical Research 111, D10202, doi:10.1029/2005JD006653.
Miao, X., Mason, J.A., Swinehart, J.B., Loope, D.B., Hanson. P.R., Goble, R.J., Liu, X.,
2007. A 10,000 year record of dune activity, dust storms, and severe drought in the
central Great Plains. Geology 35, p. 119-122.
Mason, J.A., 2001. Transport direction of Peoria Loess in Nebraska and implications for
Loess sources on the central Great Plains. Quaternary Research 56, 79-86.
Mason, J.A., Nater, E.A., Zanner, C.W., Bell, J.C., 1999. A new model of topographic
effects on the distribution of loess. Geomorphology 28, 223-236.
Mason, J.A., Swinehart, J.B., Goble, R.J., Loope, D.B., 2004. Late-Holocene dune
activity linked to hydrological drought, Nebraska Sand Hills, USA. Holocene 14,
209-217.
McCune, B., Grace, J.B., Urban, D.L., 2002. Analysis of Ecological Communities. MjM
Software Design, Gleneden Beach, OR, 284 p.
3838
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
McDonald, E.V., Busacca, A.J., 1992. Late Quaternary stratigraphy of loess in the
Channeled Scabland and Palouse regions of Washington State. Quaternary Research
38, 141-156.
Minckley T., Whitlock, C., 2000. Spatial variation of modern pollen in Oregon and
southern Washington, USA. Review of Palaeobotany and Palynology 112, 97-123.
Morrow, N., Friedl, M.A., 1998. Modeling biophysical controls on land surface
temperature and reflectance in grasslands. Agriculture and Forest Meteorology 92,
147-161.
Muhs, D.R., Bettis III, E.A., 2000. Geochemical variations in Peoria Loess of western
Iowa indicate paleowinds of midcontinental North America during last glaciation.
Quaternary Research 53, 49-61.
Mulholland, S.C., 1989. Phytolith shape frequencies in North Dakota grasses: a
comparison to general patterns. Journal of Archaeological Science 16, 489-511.
O'Geen, A.T., Busacca, A.J., 2001. Faunal burrows as indicators of paleo-vegetation in
eastern Washington, USA. Palaeogeography, Palaeoclimatology, Palaeoecology 169,
23-37.
Overpeck, J.T., Prentice, I.C., Webb III, T. 1985. Quantitative interpretation of fossil
pollen spectra: Dissimilarity coefficients and the method of modern analogs.
Quaternary Research 23, 87–108.
3939
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
PC-ORD, 1997. PC-ORD version 4 for Windows. MjM Software Design, Gleneden
Beach, OR.
Pearsall, D.M., 2000. Paleoethnobotany: A handbook of procedures. Academic Press,
San Diego, CA, 736 p.
Piperno, D.R., 1988. Phytolith analysis: an Archeological and Geological Perspective.
Academic Press, San Diego, CA, 280 p.
Piperno, D.R., 2006. Phytoliths: A comprehensive guide for archaeologists and
paleoecologists. AltaMira Press, Lanham, MD.
Richardson, C.A., McDonald, E.V., Busacca, A.J., 1997. Luminescence dating of loess
from the northwest United States. Quaternary Science Reviews (Quaternary
Geochronology) 16, 403-415.
Richardson, C.A., McDonald, E.V., Busacca, A.J., 1999. A luminescence chronology for
loess deposition in Washington State and Oregon, USA. Zeitschrift Geomorphology
116, 77-95.
Snelder, D.J., Bryan, R.B., 1995. The use of rainfall simulation tests to assess the
influence of vegetation density on soil loss on degraded rangelands in the Baringo
District, Kenya. Catena 25, 105-116.
Stevenson, B.A., Busacca, A.J., Kelly, E.F., McDonald, E.V., 2005. The stable carbon
isotope composition of soil organic carbon and pedogenic carbonates along a
4040
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
bioclimatic gradient in the Palouse region, Washington State, USA. Geoderma 124,
37-47.
Strömberg, C.A.E., 2004. Using phytolith assemblages to reconstruct the origin and
spread of grass-dominated habitats in the great plains of North America during the late
Eocene to early Miocene. Palaeogeography, Palaeoclimatology, Palaeoecology 207,
239-275.
Sweeney, M.R., Busacca, A.J., Gaylord, D.R., 2005. Topographic and climatic influences
on accelerated loess accumulation since the last glacial maximum in the Palouse,
Pacific Northwest, USA. Quaternary Research 63, 261-273.
Sweeney, M.R., Gaylord, D.R., Busacca, A.J., 2007. Evolution of Eureka Flat: A dust-
producing engine of the Palouse loess, USA. Quaternary International 162-163, 76-96.
Ter Braak, C.J.F., 1986. Canonical correspondence analysis: a new eigenvector technique
for multivariate direct gradient analysis. Ecology 67, 1167-1179.
Verma, S.D., Rust, R.H., 1969. Observations on opal phytoliths in a soil biosequence in
southeastern Minnesota. Proceedings-Soil Science Society of America 33, 749-751.
Wang, X., Qian, Y., Jiang, J., Lei, J., Zhang, W., 2003. The distribution of ephemeral
vegetation on the longitudinal dune surface and its stabilization significance in the
Gurbantunggut Desert. Acta Geographica Sinica 58, 605-609.
4141
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
Western U.S. climate historical summaries. Retrieved November 28, 2004 from Desert
Research Institute website Western Region al Climate Center. Web site:
http://www.wrcc.dri.edu/.
Whitlock, C., Sarna-Wojcicki, A.M., Bartlein, P.J., Nickmann, R.J., 2000. Environmental
history and tephrostratigraphy at Carp Lake, southwestern Columbia Basin,
Washington, USA. Palaeogeography, Palaeoclimatology, Palaeoecology 155, 7-29.
Whitmore, J., Williams, J.W., Shuman, B., Bartlein, P.J., Minckley, T., Viau, A.E., Webb
III, T., Shafer, S., Anderson, P., Brubaker, L., Gajewski, K., Sawada, M., 2005. Modern
pollen data from North America and Greenland for multi-scale paleoenvironmental
applications. Quaternary Science Reviews 24, 1828-1848.
Wilding, L. P., Drees, L. R., 1971. Biogenic opal in Ohio soils. Proceedings of the Soil
Science Society of America 37, 647-650.
Wolfe, S.A., 2001. Eolian activity. Bulletin of the Geological Survey of Canada 548,
231-240.
Zender, C.S., Bian, H., Newman, D., 2003. Mineral Dust Entrainment And Deposition
(DEAD) model: Description and 1990s dust climatology. Journal of Geophysical
Research 108, 4416, doi: 10.1029/2002JD002775.
ACKNOWLEDGEMENTS
4242
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
The authors wish to thank the St. Cloud State Diatom Herbarium for the generous use of
their facility. We also wish to thank Dr. Lewis Wixon, Dr. Matt Julius, Chad Yost, and
Bruce Wigget for their feedback and contributions to this study. This paper was
supported by NSF grant #0214508 (P.I.s: A. Busacca, C. Zender), and SCSU research
grants awarded to P. Reyerson and M. Blinnikov.
4343
916
917
918
919
920
FIGURES AND TABLES
Figure 1. Map of the Columbia Basin including vegetation zones and sample sites within
the study area. Washington vegetation data modified from Daubenmire (1972); Oregon
vegetation data modified from Gaines et al. (2004). Location information is in decimal
degrees. Mean annual temperature and mean annual precipitation data modified from
Desert Research Institute, Western U.S. Climate Historical Summaries (2004):
Marcellus (M) 47º.23N, 118º.41W MAT 9.06 ºC, MAP 301 mm
Kahlotus Ridge (K) 46º.70N, 118º.55W MAT 10.39 ºC, MAP 258 mm
Boardman (BO) 45º.65N, 119º.92W MAT 11.89 ºC, MAP 172 mm
Lindsay Prairie (L) 45º.59N, 119º.65W MAT 11.70 ºC, MAP 187 mm
4444
921
922
923
924
925
926
927
928
929
930
Table 1. Individual morphotype nomenclature used in this study and plant species they derive from. Modified in part from Blinnikov
(2005).
Abbreviation Description following International Code of Phytolith Nomenclature Taxa in Which Observed
PR Plate, short trapeziform in 3D, rectangular in 2D, epidermal short cells (ESC) Grasses
WS Short trapeziform in 3D, lobate <4 lobes on each side in 2D, sinuate margin, ESC Poa, Koeleria
WL Long trapeziform in 3D, lobate >4 lobes on each side in 2D, sinuate margin, ESC Stipa, Calamogrostis
LR Rod, very long trapeziform in 3D, mostly smooth on each side in 2D, epidermal Grasses, most heavy in long
cells (ELC) Festuca and Agropyron
LD Rod, serrated, very long trapeziform in 3D, mostly aculeate, serrated in 2D, ELC Festuca, Agropyron
SD Elongated, flat, dendriform, epidermal from seed, ELC Bromus, domesticated cereals
RO Rondel, oblong/oval: short trapeziform in 3D, oblong in top view 2D, ESC Agropyron
RP Rondel, pyramidal: truncated pyramidal in 3D, square top view 2D ESC Agropyron, Stipa
RK Rondel, keeled: short trapeziform in 3D, oval in top view 2D ESC Agropyron, Festuca, Stipa
BS Bilobate Stipa-type: trapeziform in 3D, two lobes on each side in top view, ESC Stipa, Oryzopsis
BB Bilobate: two lobes with shaft both in top and side view Stipa, Panicoid grasses
4545
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
SA Saddle: concave in top view, concave in side view, ESC Stipa (in our region)
HT Trichome: acicular/lanceolate psillate epidermal appendage (trichome) Grasses, forbs (multicell
trichomes in
Asteraceae)
HH Hair, hair base Grasses, forbs and shrubs
BL Blocky: cubic or globose, mesophyll and possibly epidermal Artemisia and other shrubs
EP Polygonal: irregular polygonal in top view, very flat on the side Forbs and shrubs
4646
220
221
222
223
224
225
226
227
Table 2. Observed total and functional groups’ vegetation cover (%, measured visually in the field) on 37 modern plots ranked by
decreasing total cover.
Plot Total Agropyron Festuca Poa Stipa3 Bromus Poaceae AsteraceaeForbs4 Shrubs
spicatum idahoensis1sandbergii2 tectorum sum sum sum sum
L4 98 15 0 3 10 40 68 2 3 25
M10 87 1 40 5 0 0 46 1 4 36
BO3 76 5 2 0 1 40 48 1 2 25
M5 75 0 10 1 0 1 12 3 5 55
K6 73 30 26 3 0 5 64 2 3 4
L1 71 10 0 3 2 50 65 5 1 0
M3 71 0 30 2 0 1 33 2 3 33
K5 70 15 45 0 0 0 60 2 3 5
K2 60 0 40 7 0 1 48 1 3 8
BO5 59 15 6 0 0 25 46 1 1 11
M1 58 0 16 2 0 1 19 5 4 30
4747
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
48
M4 57 0 50 5 0 0 55 0 2 0
K4 55 0 40 3 0 0 43 6 5 1
L5 52 10 0 5 20 10 45 3 3 1
K1 51 0 41 5 0 0 46 1 3 1
K3 51 0 30 5 0 0 35 1 3 12
L2 50 15 0 1 1 30 47 2 1 0
K7 49 25 10 5 0 0 40 1 7 1
BO1 48 5 1 0 1 10 17 0 1 30
K8 44 5 31 2 0 0 38 0 4 2
K9 44 20 12 5 0 3 40 0 3 1
L6 42 0 0 3 5 20 28 5 6 3
L3 41 0 0 2 5 20 27 2 4 8
M8 36 0 15 5 0 0 20 1 4 11
M7 35 0 20 5 0 0 25 0 2 8
M9 33 0 26 1 0 0 27 1 3 2
M2 31 0 25 1 0 0 26 0 2 3
4849
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
50
M6 30 0 20 1 0 0 21 0 3 6
BO8 27 4 5 1 6 10 26 1 0 0
K10 27 0 21 1 0 0 22 0 3 2
BO10 26 8 0 1 5 5 19 1 3 3
BO6 24 10 0 2 2 5 19 0 0 5
L7 22 1 0 2 3 10 16 2 3 1
BO4 22 3 0 0 0 10 13 1 1 7
BO9 19 6 0 1 4 5 16 1 2 0
BO2 13 6 1 0 1 2 10 1 1 1
BO7 9 1 0 1 5 0 8 1 0 0
1Also includes Festuca rubra (M9) and F. octoflora (M1, K1, K6, K8, K9, K10, BO1, BO3, BO5, BO8)
2Also includes Poa bulbosa (M3, L1, L3, L4, L6, L7) and P. juncifolia (M1)
3Stipa-group includes mainly Stipa comata, as well as S. occidentalis (BO3) and Oryzopsis hymenoides (BO7, BO8, L7)
4Also includes legumes
4951
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
52
Figure 2. Phytolith morphotype depictions. PR – Plate rectangular; WS – short
trapeziform; WL – long trapeziform; LR – long smooth trapeziform (rod); LD –
long serrated tapeziform (rod); SD – seed epidermal long cell; RO – Rondel
oblong/oval; RP – Rondel pyramidal; RK – Rondel keeled/horned; BS – Bilobate
short trapeziform Stipa-type; BB – “True” Bilobate; SA – Saddle; HT –
Trichome; HH – Hair/Hair base; BL - Blocky; EP – Epidermal non-grass. Refer to
Table 1 for the morphotype nomenclature and the list of taxa in which these forms
are observed.
Figure 3. Regression scatterplot of in field inspection vegetation cover (VC) predicted by
TOC in modern topsoil. The regression equation is VC = 20.1 + 10.7 x TOC
(R2=0.21, p=0.004). TOC is measured in percent of dry weight of extracted opal
relative to bulk soil dry weight. VC is the field-estimated percent of vegetation
ground cover for each sample.
Figure 4. Black and white example of a digital color infrared aerial image showing field
plots in Marcellus preserve. Pixel resolution is 0.16 m2, for a total of 250 pixels
per plot.
Figure 5. Regression scatterplot of NDVI values derived from the digital aerial images
predicted by total opal concentration (TOC) in topsoil. The regression equation is
NDVI = -0.432 + 0.0475 x TOC (R2=0.49, p<0.001). TOC is measured in percent
of dry weight of extracted opal relative to bulk soil dry weight. NDVI is the mean
5053
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
54
value for each plot as measured from approximately 250 pixels per plot on
multispectral aerial digital imagery.
Figure 6. Total opal concentration (TOC) and morphotype counts for sample plots. Bars
represent percentage of total. The regression equation used was VC = 20.1+10.7 x
TOC. TOC is percent of dry weight of extracted opal relative to the bulk soil dry
weight. VC is the in-field estimated percent of ground vegetation cover for each
sample. Refer to Table 1 and Fig. 2 for morphotype descriptions.
Figure 7. Canonical Correspondence Analysis scatterplot (Ter Braak, 1986; PC-ORD) of
first two axes showing position of morphotype assemblages from modern plots
and morphotypes as the main matrix and abundance of selected plant taxa as the
environmental matrix. Refer to Table1 and Fig. 2 for individual morphotype
descriptions and illustrations. Refer to Fig. 6 for morphotype percentage data per
plot.
5155
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
56
