Honors Thesis Available

HISTORICAL RECORDS OF STOMATAL INDICIES IN QUERCUS
FROM THE SOUTHEASTERN U.S.
Except where reference is made to the work of others, the work described in this thesis is
my own or was done in collaboration with my advisory committee
_______________________________
Bradford M. Cantor ‘08
Certificate of Approval:
_____________________________
______________________________
Dr. Robert A. Gastaldo, Chairman
Whipple-Coddington Professor
Department of Geology
Dr. D. Whitney King
Dr. Frank and Theodora Miselis
Professor of Chemistry
Department of Chemistry
______________________________
Dr. Bruce F. Rueger
Bermudian Palynology and Paleoecology Expert
Department of Geology
2
HISTORICAL RECORDS OF STOMATAL INDICIES IN QUERCUS
FROM THE SOUTHEASTERN U.S.
Bradford M. Cantor ‘08
A Thesis
Submitted to the Faculty of the Colby College
Department of Geology in fulfillment of the Requirements for
Honors in Geology
Waterville, ME
13 May 2008
Table of Contents
3
List of Figures and Tables
.
.
.
.
.
.
.
2
Abstract
.
.
.
.
.
.
.
.
.
3
Introduction
.
.
.
.
.
.
.
.
.
4
Materials and Methods
.
.
.
.
.
.
.
7
SD Equation 1
.
.
.
.
.
.
.
9
SI Equation 2
.
.
.
.
.
.
.
9
UI Equation 3
.
.
.
.
.
.
.
9
Results
.
.
.
.
.
.
.
.
.
10
Discussion
.
.
.
.
.
.
.
.
.
14
Summary and Conclusions
.
.
.
.
.
.
.
22
Acknowledgements
.
.
.
.
.
.
.
.
24
References
.
.
.
.
.
.
.
.
25
.
4
List of Figures and Tables
Figure 1 – Map of Auburn University Arboretum, Auburn, AL. Each species collected
appears in the key in the lower right-hand corner.
Figure 2 A – Example of stomatal count from Q. nigra (collected in 2007).
Figure 2 B – Example of stomatal count with UI measurement from Q. laurifolia
(collected in 1962).
Figure 3 – Plot of stomatal density single-area counts for all species in the present study.
Figure 4 – Plot of mean SD data (20 counts/taxon normalized to 1.00 mm2) grouped by
systematic section against pCO2.
Figure 5 A-C – Plots of normalized SI population means against pCO2 for taxa with
decreasing SI trends. These species include Q. laurifolia, Q. nigra, and Q. alba.
Figure 6 A-D – Plots of normalized SI means against pCO2 for taxa exhibiting an SI
ceiling response. These species include Q. lyrata, Q. falcata, Q. coccinea, and Q.
macrocarpa.
Figure 7 A-D – Plots of normalized SI means against pCO2 for taxa exhibiting an
anomalous SI response. These species include Q. velutina, Q. shumardii, Q. michauxii,
and Q. stellata.
Figure 7 E – Plot of normalized SI means against pCO2 for Q. nigra that exhibited a
decreasing SI trend. The last three data points collected for this species resembles
anomalous response curves.
Table 1 – List of leaves and data used for the present study.
Table 2 – Sun – Shade leaf analysis for species with high UI values.
Table 3 – Statistical results for each observed SI response.
Table 4 – Growing season temperature differences from mean temperature.
Table 5 – Palmer Drought Severity Index (PDSI) data for growing seasons of leaves
utilized in this study.
Table 6 – Precipitation data during the growing season from Auburn, AL.
5
Abstract
Modeling atmospheric carbon dioxide concentration (pCO2) is critical to
understanding the Carbon Cycle over geologic time. Recently, biological proxies have
become commonly used including stomatal frequencies on leaves. An inverse
relationship exists between epidermal stomata and the pCO2 under which the leaf grew
and expanded; this provides a record of atmospheric gases as accurate as ice core data,
another proxy, but at a higher temporal resolution. The two most commonly used leaf
parameters are Stomatal Density (SD - # of stomata/unit area) and Stomatal Index (SI the ratio between # stomata and total epidermal cells/unit area). However, stomatal
frequencies have some limitations. A CO2 “ceiling” exists where plants stop responding
to gas concentration in a linear fashion above a certain threshold. Additionally, transfer
functions calibrated from extant plants do not always correspond to fossil equivalents.
Another recent consideration is that taxonomically related plants may not exhibit similar
growth responses under the same pCO2 conditions. To test this, 12 species of Oak
(Quercus) were evaluated with respect to SD and SI over a 104 - year interval based on
trees grown under a humid subtropical climate in Lee County, AL. Herbarium specimens
dating from 1894 to the 1980s were used to supplement the historical record; a collection
made during June 2007 extended the record to the present. Herbarium samples were
taken from collections made in the same year, if available. Materials were cleared in
chromic acid, mounted on slides, after where stomata and epidermal cells were counted
using a Zeiss Axioskop and AxioVision software.
Results appear to demonstrate that Oak species respond independently to pCO2. Both
sections of Quercus, white and red, have statistically different responses to pCO2 when
results were compared. Some species, such as Q. laurifolia and Q. nigra, have similar
responses over time, but all responses tested are statistically unique. Therefore, the
present study suggests that approaches correlating data from extant taxa with data from
extinct taxa are debatable and should be considered for their qualitative observations
rather than quantitative measures.
6
Introduction
The evaluation of atmospheric proxies is necessary for further progress in
evaluating paleoclimate. Whether estimating conditions in the Holocene or the deep past,
accurate atmospheric modeling can provide a better understanding of Earth’s climate
history and predict possible changes for the future. While all atmospheric conditions are
relevant to climate modeling, carbon dioxide (CO2) concentration in the atmosphere may
be the most pertinent atmospheric gas today due to its influence on temperature and, thus,
global climate. Traditionally, Pleistocene and Holocene carbon-dioxide concentrations
(pCO2) have been estimated using air bubbles trapped in glacial ice, representing an
atmospheric snapshot from when the snow originally fell (e.g., Smith et al. 1999,
Indermuhle et al. 1999, Petit et al. 2001). In the Quaternary, glaciation at Earth’s poles
has allowed sampling ice over an interval that approaches 650 Ka (Siegenthaler et al.
2005). However, recent work has shown that while ice-core data are reliable indicators of
the climatic “big picture,” air bubbles can be biased by diffusion of trapped-gases and
glacial-sediment contamination (Wagner et al. 2002). Therefore, it is incumbent upon the
scientific community to create and develop CO2 proxies that are more accurate and
temporally resolute than ice-core data.
Several proxies other than ice-core gas measurements have been developed to
estimate pCO2 including stable-isotope relationships in paleosols and foraminiferal
boron-isotope modeling (Berner 1994, Berner and Kothavala 2001). Over the past two
decades, the plant biological proxies of stomatal index (SI - # stomata/total epidermal
cells) and stomatal density (SD - # stomata/unit area) have become increasingly popular
(e.g., Woodward 1987, Woodward and Kelly 1995, and others). These proxies rely on the
7
experimentally-determined and empirical relationship that as pCO2 increases, both SI and
SD decrease, as plants grown under these heightened CO2 levels do not need to devote as
much of their anatomy to gas exchange necessary for photosynthesis (Woodward 1987,
Royer 2001). Plants optimize these stomatal parameters between diffusion efficiency and
water conservation; the production of more stomata allows for greater gas conductance
rates, but also greater rates of evapo-transpiration (Woodward 1987, Woodward and
Bazzaz 1988, Beerling et al. 1993, Woodward and Kelly 1995). These relationships have
been documented in many taxa including, but not limited to, Quercus spp. (Kürschner
1997, Van Hoof et al. 2005, and others), Ginko biloba (Beerling et al. 1998), Alnus
glutinosa (Poole et al. 2000), Salix spp. (Rundgren and Bjorck 2003), Betula spp.
(Wagner et al. 1999, Rundgren and Bjorck 2003), and Tsuga heterophylla (Kouwenberg
et al. 2005). Calibration curves have been developed for several taxa based on the
reported inverse pCO2 – stomatal relationship. Using these calibration curves,
paleoatmospheric and paleoclimatic estimations have been made. Through these models
of ancient Earth systems, future climatic cycles can be predicted.
Stomatal frequency studies have assessed how plants responded to pCO2 in the
past with applications ranging from the Holocene to the Carboniferous (e.g., McElwain
and Chaloner 1995, Kürschner et al. 1997, Wagner 1999, Poole et al. 2000, Cleal et al.
2003). The Holocene-subfossil record consists of extant taxa whose stomatal frequencies
can be calibrated to those of living plants grown under controlled or known pCO2. SI and
SD calibration curves generated for several of these taxa have been useful to estimate
pCO2 values since the retreat of the last ice sheets approximately 12,000 YBP (Kürschner
1997, Poole et al. 2000, Beerling and Rundgren 2000, Rundgren and Bjorck 2003,
8
Rundgren et al. 2005). These provide a record that is considered to be as accurate as, and
more temporally-constrained than, ice-core data (Wagner et al. 2002). However, studies
using calibration curves into the Pre-Holocene and deep past records have extrapolated
information across broader taxonomic affinities than genus or species by employing the
nearest-living equivalent (NLE) and nearest-living relative (NLR) techniques. These
techniques are based on ecological and morphological similarities between multiple taxa
rather than their physiological relationships (McElwain and Chaloner 1995, Cleal et al.
1999, Beerling 2002, Haworth et al. 2005). While these take into consideration important
distinctions, such as similar plant morphologies and ecological niches, recent scrutiny has
shown that NLE and NLR techniques may not result in comparing the most
taxonomically similar extinct and extant taxa. Therefore, these approaches may not be as
accurate in approximating paleoatmospheric conditions (Roth-Nebelsick 2005, Cantor et
al. 2006). Presently, it has yet to be determined how taxonomically general and over what
geologic time interval stomatal frequency proxies can be applied.
This study seeks to determine how different species of the same genus, grown in
the same region under the same growth conditions, respond to increasing pCO2 over the
past century. Twelve species of Quercus (oak), a commonly fossilized genus, were
chosen and assessed to compare how similarly each oak species responded to pCO2.
There are two sub-genera within the genus Quercus -- Lobatae (red oaks) and Quercus
(white oaks) -- from which 6 species of each section were chosen for the study. Because
each taxon is unique in morphology and physiology, each Quercus species should have a
unique response to pCO2. Given each plant’s expected unique pCO2 response, intra-
9
generic similarities should result in species belonging to the same sub-genus to respond
more similarly to other members of that section.
Materials and Methods
Leaves used in the present study originate from eastern Alabama, which is located
within the Appalachian Piedmont Upland Province (Omernick 1987). All leaves were
collected from this physiographic region to control for soil, precipitation, and other
environmental factors that can affect SI and SD (Woodward 1987, Poole and Kürschner
1999). The climate of eastern Alabama is warm-temperate; most Quercus used in the
study grow in a mesophytic deciduous forest habitat, while others occur in alluvial
bottom-wetland habitats (Table 1; York 1995).
Most specimens originate from Lee County, Alabama, where many were collected
at Auburn University between 1894 and 2007 (Table 1, Figure 1). Samplings made by the
authors (1991, 2007) occurred in the Auburn University Arboretum from the exterior
canopy of each tree to ensure sun-leaf morphology. These collections were supplemented
by materials from the Auburn University Herbarium, University of Alabama Herbarium,
and the USNM Herbarium collected from Lee County or surrounding counties if more
proximal localities were not available. The Auburn University Herbarium burned twice
since the late 19th Century, limiting the historical collections available for this study.
Herbaria materials were evaluated to separate sun vs. shade leaf morphologies after
stomatal counts because shade leaves within the canopy tend to under-represent stomatal
10
index counts (Poole and Kürschner 1999). These determinations were based using the
undulation index (UI) observations of Poole and Kürschner (1999).
Sections of 1.00 cm2 were cut from the right-hand side of each leaf with the
adaxial cuticle facing upward. The sections were removed adjacent to the mid vein. In
species with trichomes, the hairs were shaved-off using an Xacto knife to allow for ease
in stomatal counting. All samples were re-hydrated in distilled water for 1 hour and then
immersed in a 10% by weight aqueous chromic acid (CrO3) solution for a duration
ranging between 12 and 24 hours (Poole and Kürschner 1999). Species with thicker
leaves were immersed for longer durations. All samples were monitored in hourly
intervals to prevent disintegration. After the dissolution of the mesophyll, the chromic
acid was decanted and the sample was water washed to neutrality. Leaf samples then
were placed under a dissecting microscope and the abaxial cuticle was removed using a
scalpel and dissecting needle. The removed cuticles were mounted onto glass slides using
BioMedia aqueous mounting gel and placed under a glass cover slip.
Cuticle mounts were photographed and analyzed using a Zeiss Axioskop and
Axiovision software. For each leaf, 20 separate 0.04 mm2 areas were counted (0.8 mm2
total) and normalized to 1.0 mm2. This differs from other studies (Wagner et al. 2002,
Van Hoof et al. 2006, and others) that have counted 7-10 areas on multiple leaves
normalized to 1.00 mm2. Stomata and epidermal cells were counted following the
methods of Poole and Kürschner (1999), and SD and SI were calculated for each sample.
Each leaf was measured according to standard sampling procedures (Poole and
Kürschner 1999) to avoid possible variation in stomatal frequencies known to occur
11
across individual leaves (Poole and Kürschner 1999). The following calculations were
used in determining SI and SD.
SD =
Eq. 1
SI =
Eq. 2
# Stomata
mm 2
# Stomata
× 100
# Stomata + # Epidermal Cells
An example of a completed leaf count has been included for reference (Fig. 2A).
Data were plotted using Grapher 4.0 software and trends over time were analyzed using
multiple linear-regression analysis without holding assumptions about the behavior of the
trends.
Climate data between 1894 and 2007 were obtained from the NOAA website,
Huntsville, AL, weather station, for the growing seasons of leaves used in this study.
Temperature, precipitation, and Palmer Drought Severity Index (PDSI) data, providing a
measure of soil moisture, also were obtained from the NOAA website, and extracted
from NetCDF format using ArcGIS software, to account for variation in stomatal
frequencies from factors that may affect SI and SD measurements on an annual basis. For
example, leaves may be biased with lower SI and SD measurements in a particularly wet
season. The archived weather data were coupled with the undulation index (UI) sinuosity
measurements to test variation and trends in stomatal frequencies against variables other
than pCO2.
Eq. 3
UI =
Ce
; C o = 2π
Co
Ae
π
Ce = circumference of the cell
Co = circumference of a circle with the same area as the cell
Ae = area of the cell
12
An example of a completed UI measurement has been included for reference (Fig.
2B).
Results
In total, 880 single-area (0.04 mm2) counts were made across 44 samples, with
each count representing one SD and one SI value normalized to 1.00 mm2. These SD
calculations ranged from 150 stomata/mm2 in Q. velutina (collected in 1934) to 1400
stomata/mm2 in Q. alba (collected in 1894). Similarly, SI calculations ranged from 4.56%
in Q. laurifolia (collected in 2007) to 20.66% in Q. alba (collected in 1894); Table 1
provides SD and SI mean values. As the range of these data indicates, both SD and SI
calculations plot over a wide array of values. In general, stomatal frequency
measurements from leaves grown under lower pCO2 conditions tend to record higher
values. This result is consistent with other stomatal-pCO2 relationships reported in the
literature (Woodward 1987, Beerling et al. 1993, van der Burgh et al. 1993, Kürschner
1997, and others).
All 20 single area (0.80 mm2 total) counts for each leaf were normalized to 1.00
mm2 and the corresponding single SD value was plotted against pCO2. These normalized
stomatal density data plot over a large range of values (Fig. 3). SD values tend to
demonstrate an inverse relationship with pCO2 as noted above, when plotted as mean
values grouped by pCO2 (i.e., mean values by year; Fig. 4). In general, the section
Quercus (white oaks) responded to increasing pCO2 with higher SD values than the
section Lobatae (red oaks). This is apparent when examining the range of values. For
13
example, Q. alba (1894 specimen in section Quercus), has the highest SD, whereas Q.
velutina (1934 specimen in section Lobatae), has the lowest SD. These general trends are
important when considered in conjunction with SI trends to complete an overall
understanding of stomatal response to changes in pCO2 over the past century. Statistical
analyses to test the significance of SD trends were not performed, as workers have
demonstrated that SD is a less reliable indicator of pCO2 than SI (Royer 2001).
Stomatal Index means of normalized data also plot over a wide range of values
(Figs. 5-7). Unlike SD, which tends to be more erratic, SI calculations reveal three
distinct trends as pCO2 has increased over the past century. These trends include: (1)
decreasing SI when plotted against increasing pCO2; (2) an apparent SI “ceiling,” or
maximum response to pCO2, as has been described in literature for other taxa (Royer
2001); and (3) anomalous responses to pCO2. It appears as if these trends are not
constrained to either section Quercus or Lobatae and occur on a species-by-species basis.
Additionally, the SI response of all normalized data collected from section Quercus is
statistically different than the response of all normalized data collected from section
Lobatae with p > 0.0001 (Table 3).
Three species exhibit a clear SI decrease in response to increasing pCO2: Quercus
laurifolia and Q. nigra (section Lobatae), and Q. alba (section Quercus; Fig 5). It is
important to note that each taxon begins the dataset with samples collected in 1927 or
earlier. This extended record provides stomatal frequency observations over a longer
temporal interval than for several other taxa used in the present study. It is possible that
these longer records reconstruct a more accurate representation of stomatal frequency
response. Regardless, these species still exhibit a few anomalous responses in particular
14
growth years (i.e., Q. laurifolia 1962, Q. nigra 1991). But, overall these taxa demonstrate
a decreasing SI trend over the past century. Although these taxa all have a decreasing SI
trend, each species shows evidence of a statistically distinct response with p < 0.0001
(Table 3).
Four species appear to exhibit what is described as a pCO2 “ceiling” response
(Royer 2001). This is one that becomes statistically weaker once a specific pCO2
threshold has been reached. In the present study, these species include: Q. coccinea and
Q. falcata, (section Lobatae), and Q. lyrata and Q. macrocarpa (section Quercus; Fig 6).
Each taxon demonstrates this response over similar values. Mean SI values range from
10.6% in Q. lyrata (collected in 2007) to 14.4% in Q. macrocarpa (collected in 1962).
When compared to the other taxa in the present study, all other values fall within this
small range, suggesting that a pCO2 ceiling has been reached. Additionally, each taxon
has an SI datum from both 1991 and 2007 that overlap statistically, further indicating that
a ceiling response is being demonstrated. The fact that data used to plot trends in Q.
macrocarpa and Q. coccinea represent only 3 years for each taxon also should be
considered. While these taxa appear to demonstrate a ceiling response, each species has a
statistically distinct response when compared to each other (p = 0.0361; Table 3).
Four species demonstrate anomalous SI responses to changes in pCO2 over the
past century. These include: Q. shumardii and Q. velutina (section Lobatae), and Q.
michauxii and Q. stellata (section Quercus; Fig 7). Quercus shumardii, Q. michauxii, and
Q. stellata all exhibit similar responses. In each taxon, the SI values for 1962 are lower
than those of 1991. Subsequently, they exhibit lower SI values in 2007 than recorded in
1991, making it appear that an SI peak for these species occurred in the late 20th Century.
15
Quercus velutina responds anomalously overall. This taxon records its lowest SI value in
1934, when pCO2 was at the lowest concentration over all years for which it was sampled
(1934-2007). Its SI response is then anomalously high in 1962, lower in 1991, and then
high again in 2007 when compared to the other taxa. Statistical tests revealed that each
species’ anomalous response was statistically significant from each other (p < 0.0001,
Table 3). Again, it is possible that if late 19th Century data were available for these taxa,
some of the noise in these trends might be resolved. However, based on the data
available, it must be assumed that other factors affected SI values for these trees. A
detailed investigation into this variation follows in the discussion.
For each leaf sampled, a sinuosity undulation index calculation was performed to
determine if one or more of the unusual leaf values might be due to data collection from a
shade leaf. There is no clear pattern to UI values that ranged from 1.07 in Q. alba
(collected in 1894) to 1.56 in Q. velutina (collected in 1934) (Table 1). Higher values
often are considered to be indicative of shade leaves, while lower values are considered
sun-leaf morphotypes. High UI in conjunction with epidermal cell density values vary
~ -30% from the species’ mean epidermal cell density are used as shade leaf evidence
(Kürschner et al. 1997). In this study, Q. velutina (collected in 1934) is considered to be a
shade leaves based on its high UI value (1.56, Table 1) coupled with a mean epidermal
cell density -26% (Table 2) than the mean value of Q. velutina collected in this study
(Table 2). All other leaves collected fall within ±30% of its species mean epidermal cell
density value and are considered to be sun leaf morphotypes.
16
Discussion
Investigators have attempted to use stomatal proxies (SI & SD) over the past two
decades to record changes in atmospheric CO2 over larger time intervals. Initially, CO2
had been tracked using SD as early as the late 19th Century (Ahrrenius 1896) and SI was
later developed to help complement and standardize the SD datasets (Salisbury 1927).
Subsequent work evaluated the accuracy of these proxies and determined that SI is a
more reliable indicator of pCO2 than SD because SD is responsive to factors other than
atmospheric gas concentration (Poole and Kürschner 1999). Although SD is not always a
reliable pCO2 indicator, SI cannot always be calculated due to preservation and
taphonomic biases (Van der Burgh et al.1993). These observations highlighted the need
for careful and methodical sampling procedures to ascertain quality paleoclimatic
approximations.
Due to the inherent questions in stomatal techniques, empirical investigations on
the relationship between SD, SI, and pCO2 have continued. Kürschner et al. (1997)
demonstrated that leaf type and sun versus shade-leaf morphotypes affect stomatal
frequency response over the past century. Poole et al. (2000) established that there is
variation in both SI and SD across a single leaf. Realities like these led to Poole and
Kürschner (1999) to propose the standardization of collection and sampling methods to
minimize any potential variation and/or error in collection.
Despite the limitations of stomatal frequency techniques, a number of studies that
resulted in estimations of pCO2 during different geologic time periods ensued. These
included the refinement of not only the Quaternary record (van der Burgh et al. 1993,
Wagner et al.1999, Rundgren and Beerling 1999), but also application of these
17
techniques into the deeper past (McElwain and Chaloner 1995, McElwain 1998). While
Quaternary samples are calibrated to the responses of modern taxa, studies dating back to
the Paleozoic employ transfer functions, known as Nearest Living Equivalent and Nearest
Living Relative (NLE, NLR) techniques, to estimate ancient pCO2 (McElwain and
Chaloner 1995, McElwain 1998). These techniques rely on morphological and ecological
similarities to draw comparisons between living and extinct taxa. After some of these
methodologies were questioned as to their reliability (Boucot and Gray 2001), work has
continued using NLE and NLR techniques with “living fossils” (Retallack 2001, Royer et
al. 2001, Beerling and Royer 2002). These studies use modern taxa like Ginkgo biloba,
which has changed little in its morphology over the Cenozoic, and compare them to their
fossil equivalents.
Other reliable pCO2 proxies emerged, including ice-core data, and were evaluated
against stomatal techniques. Several studies accurately record carbon dioxide levels in the
Holocene while at the same time ground-truthing ice-core data (Wagner 1999, McElwain
et al. 2002, van Hoof et al. 2005). In this research, workers revealed potential weaknesses
in the precision of ice-core data and revealed that annual-scale temporal resolution in
stomatal proxies (Wagner et al. 1999). Since then, several studies have focused on using
multiple species and how their congruent inverse responses record changes in pCO2
(Rundgren and Bjorck 2003, van Hoof et al. 2006).
Most recent studies have focused on applying the stomatal proxies, rather than
evaluating and refining the methodologies. Meanwhile, the results from these studies
have been incorporated into broader-scale climate models such as Geocarb (Berner and
Kothavala 2001). Recently, sources of SI variation other than pCO2 have been identified.
18
For example, it has been determined that latitude plays a role in variability within a single
genus (Garcia-Amorena et al. 2006), and leaf development at higher altitude can increase
SI values due to “thin” atmospheric conditions (McElwain 2004). Boyce (2007)
evaluated water availability as a constraint on SI, and Daly and Gastaldo (2007)
considered whether leaf orientation to sunlight during leaf expansion affects SD and SI
variability. Roth-Nebelsick’s (2005) review suggested that individual species may
respond uniquely to changes in pCO2, a finding that also has been suggested by Cantor et
al. (2006). The present study corroborates the initial hypotheses of Roth-Nebelsick
(2005) and Cantor et al. (2006). Variation across the genus Quercus is wide-spread
through all species sampled, grown under the same geographic and climatic constraints.
Systematic response is statistically distinct for all taxa over the past century of increasing
pCO2.
The present study documents different SI responses to pCO2 in the genus Quercus
for each of the 12 species sampled. This is even more significant because all the sampled
trees originated in a small geographical area that allowed for control of edaphic (soil) and
climatic conditions. Statistically significant differences between sections Lobatae and
Quercus (Table 3) indicate that their responses are unique. This is also true when
comparing all species within each and across sections (Table 3). Therefore, while it could
be expected that species from different systematic sections respond differently, it also
was revealed that species within the same taxonomic section respond distinctly.
Three SI response patterns for the genus Quercus were found in the trees from
Alabama. These responses are a decreasing SI trend, an SI ceiling, and a variety of
anomalous responses to increasing pCO2. All three also were tested to determine
19
statistically whether or not species within the trend responded congruently. Of the species
in which decreasing SI values were documented as pCO2 increased (Q. laurifolia, Q.
nigra, and Q. alba), none had a statistically similar response. Although the trends of these
taxa appear to be similar when plotted (Fig. 4), their responses are statistically distinct,
suggesting that each species responds uniquely to pCO2 over time (Table 3). Among
species exhibiting a pCO2 ceiling response, each taxon also responded uniquely (Fig. 5,
Table 3). This suggests that even when pCO2 reached 340 ppm in the early 1980s, the
concentration at which Q. robur stops responding dramatically to changes in pCO2
(Woodward and Bazzaz 1988, Royer 2001), each of the Alabama species (Q. coccinea,
Q. falcata, Q. lyrata, and Q. macrocarpa) continued to demonstrate unique responses to
pCO2 changes even after a threshold value had been reached. Finally, the species that
exhibit anomalous responses to pCO2 also reacted uniquely (Fig. 6, Table 3). While there
appears to be an SI “peak” response in the values calculated from material sampled in
1991, all species sampled exhibit statistically unique SI trends. Therefore, these species
each respond uniquely to changes in pCO2.
Several factors have been identified that potentially control variation in stomatal
response within species of Quercus. These factors include, but are not limited to
temperature, soil moisture, and precipitation (Beerling et al. 1993, Woodward and Kelly
1995, Royer 2001). Theoretically, higher temperatures during the inception of leaf
growth should result in lower SI values because they promote higher rates of evapotranspiration, forcing the plant to conserve water (Royer 2001). The growing seasons
between 1958 and 1962 in Auburn, AL, experienced extreme temperature variation
(Table 4). It experienced a colder season with up to 20% lower mean temperatures in
20
1958, while in 1962 temperatures were as much as 16% higher than mean values (Table
4). These fluctuations potentially could be responsible for the erratic SI values identified
in the dataset. However, SI values from leaves collected in 1962 range from 8.86% (Q.
nigra) to 14.45% (Q. macrocarpa) and appear low in three anomalously responding taxa
(Table 1). Yet these values only appear low when compared to the high values
representative of leaves collected in 1991. Temperature data from 1991 does not suggest
a great aberration from the norm and, therefore, temperature is likely not responsible for
statistical variation observed in SI (Table 4).
Soil moisture is another environmental factor that has been shown to cause SD
variation (Woodward and Kelly 1995, Royer 2001). In an examination of soil moisture in
the Auburn area using the Palmer Drought Severity Index (PDSI) acquired from NOAA,
there are several years in the present study’s dataset with extremely wet and extremely
dry soil conditions. Dry growing seasons (February to April) include 1927 and 1934
(Table 5). Quercus velutina, the only species sampled from 1934, had an anomalously
low SI value (Fig. 6, Table 1). Although this particular sample was considered a shade
leaf, it is possible that shade leaf SI values were exaggerated due to drought conditions
(Tables 2, 5). PDSI appears to have no affect on Q. laurifolia (collected in 1927), as this
sample provides a datum that is consistent with the species’ overall decreasing trend.
Conversely, 1962 was a particularly wet year and PDSI data reveal that soil moisture was
extremely high. However, no stomatal parameters from leaves collected in 1962 appear to
represent extremely abnormal data (Table 1, Figs. 4, 5, 6). Several SD values appear to be
high (i.e., Q. falcata), which is consistent with the reported results on SD sensitivity (i.e.,
21
Royer 2001). Therefore, the consistency within the SI data seems to indicate that this
stomatal proxy is not greatly affected by soil moisture in this dataset.
Supplementing the PDSI data are precipitation data, which provide an indication
of humidity and climatic conditions during the past century in the southeastern U.S.
Leaves growing in humid and rainy conditions are expected to have higher SI values,
because water-loss due to evapo-trainspiration will not be as significant (Woodward and
Kelly 1995). April 1911 was extremely wet with a 204% increase in precipitation (Table
5), but the two species sampled (Q. lyrata and Q. laurifolia) record values that are
consistent with the other data collected for those taxa. Precipitation in 1991 was ~ 60%
higher than average for each month during the growing season. This is intriguing because
there is an observed SI peak in 1991 for three taxa grouped into those that show an
anomalous response and one taxon placed into the decreasing SI trend group (Fig. 6).
Although these species each exhibit statistically distinct responses over time, it must be
recognized that all had high SI values in 1991. Heavy rainfall in that year is a likely
explanation for these SI values. This is especially pertinent because it complements
recent work indicating that leaf venation density and, hence, other anatomical features of
leaves may be a function of water availability as well as pCO2 (Boyce 2007).
The present study documented that 12 species of oak, all collected within a
constrained physical region, each had distinct responses to increasing pCO2 over the past
century. This result confirms initial hypotheses that have suggested species-level
variation for SI (Roth-Nebelsick 2005, Cantor et al. 2006). The relationship between any
one species and pCO2 is taxon-dependent and is not consistent within even a single genus.
This insight is important for studies that have used similar decreasing SI trends either in
22
two species (van Hoof et al. 2006) or two different genera (Rundgren and Beerling 1999)
to extrapolate past pCO2 values because their relationship might not be as straightforward as was once believed (Cleal et al. 1999). Additionally, studies using NLE
(McElwain and Chaloner 1995) and NLR (Beerling and Royer 2002, and others)
techniques, including the use of “living fossils” to extrapolate pCO2 back as far as the
Paleozoic, should be recognized as providing more qualitative than quantitative trends
because even genus-level assumptions about how SI values relate to atmospheric gas
concentration are speculative. Studies incorporating SI-based proxies into global climate
models should use these data only to supplement findings rather than factor them into the
equation (Berner and Kothavala 2001).
Instead, a potential technique for tracking changes in atmospheric gas
concentration in the recent and deep past could employ the use of SI and SD
measurements through a stratigraphic section. SD and SI changes observed upsection
(forward in time) suggest changing climatic conditions. These measurements then could
be referenced against independent quantitative observations, such as oxygen or
foraminiferal boron isotopes (Berner and Kothavala 2001) to confirm the atmospheric
variation. But, even here, it must be noted that the taphonomic conditions that promote
leaf preservation are incompatible with the genesis of soil carbonate nodule production in
many instances (Dimichele and Gastaldo 2008). Because SI and SD are more temporally
resolute than other paleoclimate proxies (Wagner et al. 2002), rapid fluctuations in
stomatal frequencies throughout a stratigraphic section might reflect variable climate
during a warming trend. Paleoclimatic proxies should be used collectively and not as
stand-alone techniques (Poole et al. 2000).
23
Studies that have attempted to record changes in one species throughout the
Quaternary likely provide for an accurate pCO2 approximation over time as long as
calibration curves are constructed with appropriately large datasets and specimen
collection occurs within a consistent depositional environment. If changes in SI and SD
can be correlated to other climatic signatures, such as ice-core data and parasequences, a
complementary dataset can be constructed to better understand Earth processes. It could
be possible to extrapolate well-approximated pCO2 conditions from the Quaternary to the
deeper past if the assumptions are based on known mechanisms like Milankovic cycles.
Because Milankovic cyclicity patterns today are well-known and ice-core data,
parasequences, and stomatal proxies can be measured throughout the Quaternary, data
aberrations recorded in these proxies from the deeper past can suggest differential
cyclicity and different mechanisms controlling climate.
More creative methods to quantitatively approximate pCO2 using known
calibration curves for individual species should be considered. For example, a multispecies fossil assemblage in which there are several representatives of a single genus
could be averaged based on the aggregate of all individual calibration curve responses.
For a single genus, some species might over-represent the generic response to pCO2 while
some species might under-represent the response to pCO2. If the average of all data used
in this study is taken, the approximate CO2 value would likely fall within an acceptable
error range of the ambient concentration. These approximations would be extremely
precise, but at the present time, it is not known how many species are necessary to
document in order to establish an accurate quantitative approximation.
24
Qualitative observations based on fossil-assemblage averages could be used as
well. For example, if due to some taphonomic fortuitous chance all leaves in this study
were time-averaged to one year and preserved in a fossil assemblage for study 100
million years from now, the same responses observed in this study would be recorded in
the future. As was observed, some taxa might decrease in SI, some exhibit a ceiling
response, while others are anomalous as samples are observed upsection. These results
observed in a stratigraphic context would suggest increasing pCO2, with some anomalous
climatic conditions as was observed in this study. If another section was observed that all
species in the assemblage were averaged to a ceiling response, it might be concluded that
the quantity of CO2 in the atmosphere was enough to not stress the growing plants. If all
species were averaged to a decreasing response to pCO2, rapidly changing climatic
conditions would be implied. Correlated sections across a large basin could record
differential paleo-latitude within the same species or other changes in factors intrinsic to
SI and SD variability. Therefore, while there are certainly limits on the conclusions that
can be drawn from one species’ change in SI and SD over time, a multi-species approach
to compliment sequence stratigraphic data could prove to be extremely useful in
estimating paleoclimate.
Summary and Conclusions
The present study observed SI response to increasing levels of pCO2 over the past
century in 12 species of oak from a small geographic region. Three responses were
observed. These included decreasing SI, SI ceiling, and anomalous SI responses to pCO2.
While many species exhibited similar trends, no one trend was statistically similar to
25
another. Each species had a distinct and unique response to pCO2. Therefore, even
species within the same genus may not respond to changing pCO2 as similarly as
previously thought.
This result has implications for future stomatal proxy research. This study
confirms the initial hypotheses that studies using NLE, NLR, living fossil equivalent, and
other extrapolations from living to extinct species are questionable because living species
even within the same genus do not respond similarly to changing pCO2. It might also be
possible to average SI values for all leaf fossils from the same genus in an assemblage to
account for all variation across one genus. This approximation might place a quantitative
value on pCO2 to within a small margin of error. This hypothesis needs to be tested and it
is uncertain how many species from the same genus would need to be included to obtain
an accurate representation of all SI variation.
Additionally, creative techniques should be attempted to correlate stomatal
proxies with other climate proxies and sequence stratigraphy. A qualitative stomatal
dataset can help supplement independent quantitative results and should not be
discounted. Investigation into how SI and SD in living taxa have responded to changing
pCO2 in the Quaternary could be valuable because an intra-generic response to Earth
processes such as Milankovic cycles could potentially be compared to intra-generic
responses to Earth processes in deep time.
26
Acknowledgements
There are many people who have contributed to this project that I would like to
thank. I would like to thank each of the herbarium curators and arboretum botanists that
helped with this work including Curtis Hansen, curator of the Auburn University
Herbarium, Steve Ginzbarg, curator of the University of Alabama herbarium, and
Deborah Bell, curator of the U.S.N.M. herbarium in Washington. Additional climate-data
collection was made possible through the help of Kris White, the climatological data
team leader at NOAA’s station in Huntsville, AL. Data extraction using ArcGIS software
was aided by Manuel Gimond, of Colby College Academic ITS. Special thanks go out to
Liam O’Brien, statistics professor at Colby College, for his help running statistical
analyses on this project. I would also like to thank Edward Yeterian, Dean of Faculty at
Colby College, and the Colby College Department of Geology for monetary support to
attend national conferences of the Geological Society of America. Finally, I would like to
thank my advisor Robert A. Gastaldo for his support, insight, criticism, and overall help
while working on this project.
27
References
Ahrrenius, S., 1896, On the influence of carbonic acid in the air on the temperature on the
ground: Philosophical Magazine, v. 41, p. 237-276.
Beerling, D.J., 2002, Low atmospheric CO2 levels from the Permo-Carboniferous
Glaciation inferred from fossil lycopsids: PNAS, v. 99, p. 12567-12571.
Beerling, J.D., Chaloner, W.G., Huntley, B., Pearson, J.A., Tooley, M.J., 1993, Stomatal
Density Responds to the Glacial Cycle of Environmental Change: Proceedings:
Biological Sciences, v. 251, p. 133-138.
Beerling, D.J., McElwain, J.C., Osborne, C.P., 1998, Stomatal responses of the ‘living
fossil’ Ginkgo biloba L. to changes in atmospheric CO2 concentrations: Journal of
Experimental Botany, v. 49, p. 1603-1607.
Beerling, D.J., Royer, D.L., 2002, Fossil plants as indicators of the Phanerozoic global
carbon cycle: Annual Reviews in Earth and Planetary Sciences, vol. 30, p. 526556.
Beerling, D.J., Rundgren, M., 2000, Leaf metabolic and morphological responses of
dwarf willow (Salix herbacia) in the sub-Arctic to the past 9000 years of global
environmental change: New Phytologist, v. 145, p. 257-269.
Berner, R.A., 1994, GEOCARB II; A revised model of atmospheric CO2 over
Phanerozoic time: American Journal of Science, v. 294, p. 56-91.
Berner, R.A., Kothavala, Z., 2001, Geocarb III: A revised model of Atmospheric CO2
over phanerozoic time: American Journal of Science, v. 301, p. 182-204.
Boucot, A.J., Gray, J., 2001, A critique of Phanerozoic climatic models involving
changes in the CO2 content of the atmosphere: Earth-Science Reviews, v. 56, p. 1159.
Boyce, C.K., 2007, Seeing the forest with the leaves—clues of canopy placement from
leaf venation characteristics: Geological Society of America, Abstracts with
Programs, v. 39, no. 6, p. 23.
Cantor, B.M., Aigler, B.V., Pace, D.W., Reid, S.B., Thompson, C.Y., Gastaldo, R.A.,
2006, Intra- and interspecific variation in stomatal proxies for Quercus and Nyssa
28
in the subtropical Southeastern U.S.: Geological Society of America, Abstracts
with Programs, v. 38, p. 487.
Cleal, C.J., James, R.M., Zodrow, E.L., 1999, Variation in stomatal density in the Late
Carboniferous gymnosperm frond Neuropteris ovata: Palaios, v. 14, p. 180-185.
Daly, R.G., Gastaldo, R.A., 2007, The effect on stomatal proxies as a function of
geographical position relative to growing season sunlight: Geological Society of
America, Abstracts with Programs, v. 39, p. 301.
DiMichele, W.A., Gastaldo, R.A., 2008, Plant Paleoecology in Deep Time: Annals of the
Missouri Botanical Gardens, v. 95, no. 1, p. 144-198.
Garcia-Amorena, I., Wagner, F., van Hoof, T.B., Manzaneque, F.G., 2006, Stomatal
responses in deciduous oaks from southern Europeto the anthropogenic
atmospheric CO2 increase; refining the stomatal-based CO2 proxy: Review of
Palaeobotany and Palynology, v. 141, p. 303-312.
Haworth, M., Hesselbo, S.P., McElwain, J.C., Robinson, S.A., Brunt, J.W., 2005, MidCretaceous pCO2 based on stomata of the extinct conifer Pseudofrenelopsis
(Cheirolepidiaceae): Geology, v. 33, p. 749-752.
Indermuhle, A., Stocker, T.F., Joos, F., Fischer, H., Smith, H.J., Wahlen, M., Deck, B.,
Mastroianni, D., Tschumi, J., Blunier, T., Meyer, R., Stauffer, B., 1999, Holocene
carbon cycle dynamics based on CO2 trapped in ice at Taylor Dome, Antarctica:
Nature, v. 398, p. 121-126.
Kouwenberg, L., Wagner, R., Kürschner, W.M., Visscher, H., 2005, Atmospheric CO2
fluctuations reconstructed by stomatal frequency analysis of Tsuga heterophylla
needles: Geology, v. 33, p. 33-36.
Kürschner, W.M., 1997, The anatomical diversity of recent and fossil leaves of the
durmast oak (Quercus petraea Lieblein/ Q. pseudocastanea Goeppert)implications for their use as biosensors of paleoatmospheric CO2 levels: Review
of Palaeobotany and Palynology, v. 96, p. 1-30.
McElwain, J.C., 1998, Do fossil plants signal paleoatmospheric CO2 concentration in the
geologic past?: The Royal Society, v. 353, p. 83-96.
McElwain, J.C., 2004, Climate-independent paleoaltimetry using stomatal density in
fossil leaves as a proxy for CO2 partial pressure: Geology, v. 32, p. 1017-1020.
McElwain, J.C., Chaloner, W.G., 1995, Stomatal density and index of fossil plants track
atmospheric carbon dioxide in the Paleozoic: Annals of Botany, v. 76, p. 389-395.
29
McElwain, J.C., Mayle, F.E., Beerling, D.J., 2002, Stomatal evidence for a decline in
atmospheric CO2 concentration during the Younger Dryas stadial: a comparison
with Antarctic ice core records: Journal of Quaternary Science, v. 17, p. 21-19.
Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.M., Basile, I., Bender, I.,
Chappellaz, J., Davis, M., Delaygue, G., Delmott, M., Kotlyakov, V.M., Legrand,
M., Lipenkov, V.Y., Lorius, C., Pepin, L., Ritz, C., Satzman, E., Stievenard, M.,
1999, Climate and atmospheric history of the past 420,000 years from the Vostok
ice core, Antarctica: Nature, v. 399, p. 429-426.
Poole, I., Lawson, T., Weyers, J.D.B., Raven, J.A., 2000, Effect of elevated CO2 on the
stomatal distribution and leaf physiology of Alnus glutinosa: New Phytologist, v.
145, p. 511-521.
Poole, I., Kürschner, W.M., 1999, Stomatal density and index; the practice, in Jones,
T.P., Rowe, N.P., ed., Fossil Plants and Spores: modern techniques. Geological
Society, London, 251-256.
Omernick, J.M., 1987, Ecoregions of the conterminous United States: Annals of the
Association of American Geographers, v. 77, p. 118-125.
Retallack, G.J., 2001, A 300-million year record of atmospheric carbon dioxide from
fossil plant cuticles: Nature, v. 411, p. 287-289.
Roth-Nebelsick, A., 2005, Reconstructing atmospheric carbon dioxide with stomata:
possibilities and limitations of a botanical pCO2-sensor: Trees – Structure and
Function, v. 19, p. 251-265.
Royer, D.L., 2001, Stomatal density and stomatal index as indicators of paleoatmospheric
CO2 concentration: Review of Paleobotany and Palynology, v. 114, p. 1-28.
Royer, D.L., Berner, R.A., Beerling, D.J., 2001, Phanerozoic atmospheric CO2 change:
evaluating geochemical and paleobiological approaches: Earth Science Reviews,
v. 54, p. 349-392.
Rundgren, M., Beerling, D.J., 1999, A Holocene CO2 record from the stomatal index of
subfossil Salix herbacea L. leaves from northern Sweden: The Holocene v. 9,5
pp. 509-513.
Rundgren, M., Björck, S., 2003, Late glacial and early Holocene variations in
atmospheric CO2 concentration indicated by high-resolution stomatal index data:
Earth and Planetary Science Letters, v. 213, p. 191-204.
Rundgren, M., Björck, S., Hammarlund, D., 2005, Last interglacial atmospheric CO2
changes from stomatl index data and their relation to climate variations: Global
and Planetary Change, v. 49, p. 47-62.
30
Salisbury, E.J., 1927, On the causes and ecological significance of stomatal frequency,
with special reference to the woodland flora: Philosophical Transactions of the
Royal Society of London v. 216 p. 1-65.
Smith, R.C., Ainley, D., Baker, K., Domack, E., Emslie, S., Fraser, B., Kennett, J.,
Leventer, A., Mosley-Thompson, E., Stammerjohn, S., Vernet, M., 1999, Marine
ecosystem sensitivity to climate change: Bioscience, v. 49, p. 393-404.
Siegenthaler, U., Stocker, T.F., Monnin, E., Luthi, D., Schwander, J., Stauffer, B.,
Raynaud, D., Barnola, J.M., Fischer, H., Masson-Delmotte, V., Jouzel, J., 2005,
Stable carbon cycle-climate relationship during the Late Pleistocene: Science,
v. 310, p. 1313-1317.
Van Der Burgh, J., Visscher, H., Dilcher, D.L., Kürschner,W.M., 1993, Paleoatmospheric
Signatures in Neogene Fossil Leaves: Science, v. 260, p. 1788-1790.
Van Hoof, T.B., Kaspers, K.A., Wagner, F., Roderick S.W., Kürschner, W.M., Visscher,
H., 2005, Atmospheric CO2 during the 13th century AD: reconciliation of data
from ice core measurements and stomatal frequency analysis: Tellus B, v. 47.,
p. 351-355.
Van Hoof, T.B., Kürschner, W.M., Wagner, F., Visscher, H., 2006, Stomatal index
response of Quercus robur and Quercus petraea to the anthropogenic atmospheric
CO2 increase: Plant Ecology, v. 183, p. 237-247.
Wagner, F., Bohncke, S.J.P., Dilcher, D., Kürschner, W.M., Van Geel, B., Visscher, H.,
1999, Century-scale shifts in Early Holocene atmospheric CO2 concentration:
Science, v. 285, p. 1971-1973.
Wagner, F., Aaby, B., Visscher, H., 2002, Rapid atmospheric CO2 changes associated
with the 8,200-years-B.P. cooling event: PNAS, v. 99, p. 12011-12014.
Woodward F.I., 1987, Stomatal numbers are sensitive to increases in CO2 from preindustrial levels. Nature v. 327 617-618.
Woodward, F.I., Bazzaz, F.A., 1988, The responses of stomatal density to CO2 partial
pressure: Journal of Experimental Botany, v. 39, p. 1771-1781.
Woodward, F.I., Kelly, C.K., 1995, The influence of CO2 concentration on stomatal
density: New Phytologist, v. 131, p. 311-327.
York, Harlan H., 1995, 100 Forest trees of Alabama, 2nd ed.: Alabama Forestry
Comission and Division of Vocational Education, Alabama State Dept. of
Education, 111 p.
A
B
C
D
E
1
2
3
4
Q. laurifolia - C3
Q. falcata- A2
Q. coccinea- C3
Q. michauxii - B3
Q. macrocarpa - D3
Q. lyrata - B4
Q. alba - C3
Key
Q. nigra- A3
Q. muhlenbergii - A2
White Oaks
Q. stellata- C3
Q. shumardii - D2
Red Oaks
Q. velutina- D2
Figure 1. Map of the Donald E. Davis Arboretum at Auburn University, Auburn, AL. This figure was
modified from the map at http://www.auburn.edu/arboretum/images/content/map/arb-map-06.gif.
The location of each red and white oak taxon used for this study is indicated.
A
B
Figure 2. Examples of stomatal counts. (A) Stomatal count from Q. nigra collected in 2007. (B) Stomatal count with undulation index
2
(UI) measurement of epidermal cell from Q. macrocarpa collected in 1962. Both areas represent 0.04 mm .
Figure 3. Plot of normalized single-area counts for SD against pCO2, where values range from
2
150 to 1400 stomata/mm . Many SD values overlap. Overall, SD tends to decrease as pCO2
increases, but data are widely distributed.
Figure 4. Plot of mean SD data (20 counts/taxon normalized to 1.00 mm2) grouped by systematic section against pCO2.
Each datum represents mean values for at least 20 single-area counts with 95% confidence interval error bars drawn. In
years for which multiple leaves were obtained, the n-value is higher (Table 2). A general decreasing SD trend is observed as
pCO2 increases. White oak taxa tend to plot with higher SD values than red oak taxa.
Figure 5. Plots of normalized SI population means against pCO2 for taxa with decreasing SI trends. Each datum
represents mean values for 20 single-area counts with 95% confidence interval error bars drawn and these taxa represent
leaf samples from ≥4 years. (A, B) Quercus laurifolia and Q. nigra, section Lobatae, exhibit a decreasing SI trend as
pCO2 increases. (C) Quercus alba, section Quercus, also exhibits a decreasing SI trend as pCO2 increases, but represents
higher SI values. The difference in responses between red oaks and white oaks is significantly different statistically.
Figure 6. Plots of normalized SI means against pCO2 for taxa exhibiting an SI ceiling response. Each datum
represents mean values for 20 single-area counts with 95% confidence interval error bars drawn. (A, B) Quercus
lyrata, section Quercus, and Q. falcata, section Lobatae, demonstrate a ceiling response as their SI responses
become less dramatic as pCO2 increases. (C, D) Quercus coccinea, section Lobatae, and Q. macrocarpa, section
Quercus, represent SI values that slightly decrease as pCO2 increases. In all four taxa, the SI measurements for
pCO2 at 354 ppmv and 385 ppmv (1991 and 2007, respectively) overlap statistically.
Figure 7. Plots of normalized SI means against pCO2 for taxa exhibiting anomalous SI responses to pCO2. Each datum represents mean
values for 20 single-area counts with 95% confidence interval error bars. (A) Quercus velutina, section Lobatae, demonstrates anomalous
SI responses with increasing SI values over time. SI values in 1934 plot lower than SI values from 1958; and 1991 plot lower than 2007
SI values. (B, C, D) Quercus shumardii, section Lobatae, and Q. macrocarpa and Q. stellata, section Quercus, respond anomalously to
pCO2. Some mean values overlap statistically (i.e., 6 B,C), but inconsistent trends are recorded overall. It is possible that these plots
mimic other trends recorded: either ceiling responses or decreasing trends, but because of low sample numbers and a limited temporal
record, further extrapolation is not possible.
Taxon
Q. coccinea
Q. coccinea
Q. coccinea
Q. falcata
Q. falcata
Q. falcata
Q. falcata
Q. laurifolia
Q. laurifolia
Q. laurifolia
Q. laurifolia
Q. laurifolia
Q. laurifolia
Q. nigra
Q. nigra
Q. nigra
Q. nigra
Q. nigra
Q. shumardii
Q. shumardii
Q. shumardii
Q. velutina
Q. velutina
Q. velutina
Q. velutina
Section
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Lobatae (Red)
Years Sampled
1958
1991
2007
1924
1962
1991
2007
1911
1927
1953
1962
1991
2007
1924
1953
1962
1991
2007
1962
1991
2007
1934
1963
1991
2007
Reported Habitat
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Stream and Wetland
Stream and Wetland
Stream and Wetland
Stream and Wetland
Stream and Wetland
Stream and Wetland
Stream and Wetland
Stream and Wetland
Stream and Wetland
Stream and Wetland
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Location Sampled
AUC
AUA
AUA
AUC
AUC
AUA
AUA
AHC
SBHC
AUC
AUC
AUA
AUA
AU
AUA
LCA
AUA
AUA
LCA
AUA
AUA
AU
MCA
AUA
AUA
Fresh/Historical
H - AU
F
F
H - AU
H - AU
F
F
H - USNM
H - AU
H - AU
H - AU
F
F
H - AU
H - AU
H - AU
F
F
H - AU
F
F
H - AU
H - AU
F
F
N
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
Mean SI
13.17%
11.52%
11.15%
12.11%
12.03%
11.21%
11.14%
11.30%
10.77%
8.74%
9.44%
5.90%
6.33%
10.86%
9.12%
8.86%
9.81%
7.20%
9.13%
11.68%
11.42%
6.90%
10.79%
7.55%
9.25%
Standard Deviation Mean SD
1.00%
811
0.75%
736
0.99%
595
1.18%
716
0.64%
845
1.18%
574
0.92%
758
1.23%
585
1.60%
564
0.56%
376
1.06%
576
0.77%
420
0.96%
351
1.22%
511
0.83%
420
0.77%
474
1.23%
713
0.75%
335
0.83%
394
1.80%
461
1.10%
648
0.86%
243
1.09%
719
0.85%
325
1.08%
339
UI
1.35
1.33
1.32
1.20
1.40
1.16
1.37
1.32
1.23
1.18
1.34
1.32
1.39
1.29
1.33
1.37
1.27
1.17
1.42
1.19
1.20
1.56
1.36
1.55
1.35
Table 1. List of leaves used for the present study. Leaves are organized by year alphabetically by section. Habitat descriptions are derived from York
(1953). Sampled locations include: Auburn University Campus, Auburn, AL (AUC); Auburn University Arboretum, Auburn, AL (AUA); St. Bernard
Historical Collection (SBHC); Auburn, AL (AU); Lee County, AL (LCA); Macon County, AL (MCA); Abbevile, Henry County, AL (AHC); and
Tuskeegee University, Tuskeegee, AL (TU). All fresh leaves were collected from the Auburn University Arboretum. Leaves sampled from historical
collections come from three different herbaria: Auburn University Herbarium (H - AU); United States National Museum Herbarium (H - USNM); and the
University of Alabama Herbarium (H - UA).
Taxon
Q. alba
Q. alba
Q. alba
Q. alba
Q. lyrata
Q. lyrata
Q. lyrata
Q. lyrata
Q. macrocarpa
Q. macrocarpa
Q. macrocarpa
Q. michauxii
Q. michauxii
Q. michauxii
Q. muhlenbergii
Q. muhlenbergii
Q. stellata
Q. stellata
Q. stellata
Section
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Quercus (White)
Years Sampled
1897
1960
1991
2007
1911
1962
1991
2007
1962
1991
2007
1962
1991
2007
1991
2007
1953
1991
2007
Reported Habitat
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Low Flatwoods
Low Flatwoods
Low Flatwoods
Low Flatwoods
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Stream and Wetland
Stream and Wetland
Stream and Wetland
Low Flatwoods
Low Flatwoods
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Eastern Mesophytic Deciduous
Location Sampled Fresh/Historical
AU
H - UA
LCA
H - AU
AUA
F
AUA
F
AHC
H - USNM
AUC
H - AU
AUA
F
AUA
F
TU
H - AU
AUA
F
AUA
F
AUC
H - AU
AUA
F
AUA
F
AUA
F
AUA
F
AUC
H - AU
AUA
F
AUA
F
N
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
Mean SI Standard Deviation Mean SD
18.11%
1.58%
950
15.07%
1.19%
1015
12.51%
1.20%
764
10.88%
1.49%
446
12.12%
0.98%
830
12.01%
0.85%
805
11.02%
1.73%
720
10.62%
1.11%
669
14.45%
1.44%
1044
12.57%
1.36%
799
11.92%
0.79%
700
12.81%
1.23%
693
13.38%
1.61%
709
9.91%
1.03%
574
13.58%
1.05%
763
13.05%
1.19%
818
9.80%
1.01%
411
10.88%
1.78%
538
8.69%
0.65%
435
UI
1.07
1.20
1.22
1.17
1.27
1.15
1.35
1.22
1.41
1.24
1.29
1.21
1.51
1.27
1.49
1.04
1.19
1.47
1.23
Table 1. List of leaves used for the present study. Leaves are organized by year alphabetically by section. Habitat descriptions are derived from York
(1953). Sampled locations include: Auburn University Campus, Auburn, AL (AUC); Auburn University Arboretum, Auburn, AL (AUA); St. Bernard
Historical Collection (SBHC); Auburn, AL (AU); Lee County, AL (LCA); Macon County, AL (MCA); Abbevile, Henry County, AL (AHC); and
Tuskeegee University, Tuskeegee, AL (TU). All fresh leaves were collected from the Auburn University Arboretum. Leaves sampled from historical
collections come from three different herbaria: Auburn University Herbarium (H - AU); United States National Museum Herbarium (H - USNM); and the
University of Alabama Herbarium (H - UA).
Species
Q. falcata 1962
Q. velutina 1934
Q. velutina 1991
Q. stellata 1991
Q. laurifolia 1991
Q. laurifolia 2007
Undulation Index Species Mean
1.4
191
1.56
177
1.55
167
1.47
167
1.32
193
1.39
206
Year Mean
247
131
159
177
271
207
% Difference Sun/Shade?
29%
Sun
-26%
Sun/Shade
-5%
Sun
6%
Sun
40%
Sun
0%
Sun
Table 2. Sun-Shade leaf analysis of species with high UI. After UI and epidermal cell density analysis of these
taxa, only one shade leaf was revealed. Quercus velutina (collected in 1934) is on the borderline of sun/shade leaf
morphology because it has a high UI and epidermal cell density approaching 30% different from the mean
epidermal cell density for Q. velutina collected in the present study.
Test
Statistical Significance
Section Lobatae
Section Quercus
Lobatae v. Quercus
Decreasing SI
SI Ceiling
Anomalous SI
p < 0.0001
p = 0.0296
p < 0.0001
p < 0.0001
p = 0.0361
p < 0.0001
Table 3. Statistical results for each observed SI response.
Comparisions of taxa within sections Lobatae and Quercus were
performed in conjunction with cross-section comparisions. Analyses
comparing the responses of species exhibiting a decreasing response,
SI ceiling response, and anomalous SI responses were performed as
well. In every analysis (6 total) there was a statistically significant
Year
February
March
April
1911
1924
1927
1934
1950
1958
1960
1962
1963
1991
2007
10.22%
3.28%
14.42%
5.11%
8.21%
18.80%
10.04%
15.69%
7.30%
4.20%
5.47%
4.26%
8.99%
2.05%
1.10%
-4.73%
10.73%
20.66%
11.67%
-9.31%
2.05%
-14.98%
-2.75%
0.27%
4.81%
-0.82%
-6.18%
-0.69%
-5.77%
-4.40%
-3.57%
1.10%
3.02%
Table 4. Growing season temperature differences from mean
temperature. From 1958 to 1962 there are large temperature
variations. However, the corresponding SI values do not show
abnormal data for these years, suggesting temperature has little
affect on SI variation.
Year
Month
PDSI
Soil Conditions
1894
1894
1894
1911
1911
1911
1924
1924
1924
1927
1927
1927
1934
1934
1934
1953
1953
1953
1958
1958
1958
1960
1960
1960
1962
1962
1962
1963
1963
1963
1991
1991
1991
February
March
April
February
March
April
February
March
April
February
March
April
February
March
April
February
March
April
February
March
April
February
March
April
February
March
April
February
March
April
February
March
April
-0.7
-1.4
-1.8
-2.1
-3.0
-1.9
-0.1
-0.8
0.4
-1.1
-1.1
-1.4
-4.5
-3.8
-4.0
-0.8
-0.7
-1.3
-0.8
-1.0
-0.8
0.4
0.7
-0.8
3.7
3.1
3.2
-1.6
0.1
0.4
0.2
0.2
0.6
NORMAL
DRY
DRY
DRY
DRY
DRY
NORMAL
NORMAL
NORMAL
DRY
DRY
DRY
DRY
DRY
DRY
NORMAL
NORMAL
DRY
NORMAL
DRY
NORMAL
NORMAL
NORMAL
NORMAL
WET
WET
WET
DRY
NORMAL
NORMAL
NORMAL
NORMAL
NORMAL
Table 5. Palmer Drought Severity Index (PDSI) data for
growing seasons of leaves collected in this study. In
1934 there was a severe drought, possibly exaggerating
shade-leaf SI values observed in Q. velutina. Very wet
conditions occurred in 1962, but normal SI values over
this time span indicate high PDSI values have little
affect on these leaves.
Year
February
March
April
% Difference February
% Difference March
% Difference April
1894
1911
1924
1927
1934
1950
1958
1960
1962
1963
1991
2007
Montly Mean (1894-2007)
4.98
5.19
5.35
3.93
3.84
7.36
3.99
2.58
8.38
1.88
8.46
2.53
4.73
3.83
2.05
5.67
6.04
8.4
7.28
2.79
8.31
6.26
10.46
9.33
1.12
5.85
3.05
14
7.87
3.81
3.15
2.88
8.84
2.09
4.57
9.05
10.29
3.28
4.61
5%
10%
13%
-17%
-19%
56%
-16%
-45%
77%
-60%
79%
-47%
-35%
-65%
-3%
3%
44%
24%
-52%
42%
7%
79%
59%
-81%
-34%
204%
71%
-17%
-32%
-38%
92%
-55%
-1%
96%
123%
-29%
Table 6. Precipitation data during the growing season from Auburn, AL. The wettest season occurred in 1991, when four taxa
exhibited an SI peak over this time period (Fig. 6). It is likely that water availability affected SI values and venation for these leaves as
suggested by Boyce (2007).