Southeastern Naturalist
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
138
2014 SOUTHEASTERN NATURALIST 13(1):138–155
Ground-dwelling Beetle Responses to Long-term
Precipitation Alterations in a Hardwood Forest
Ray S. Williams1,*, Bryan S. Marbert1,2, Melany C. Fisk1,3, and Paul J. Hanson4
Abstract - It is widely predicted that regional precipitation patterns may be altered due
to climate change, and these changes may affect areas with extensive forests. Therefore,
studies investigating the role of this climate driver on forest floor fauna are timely. We
examined the impact of precipitation alteration over 13 years on Coleoptera (specifically
Family Carabidae) communities in a temperate forest by testing the effects of dry (33%
precipitation interception), ambient (control), and wet (33% precipitation addition) treatments.
We collected insects in pitfall traps and quantified forest-floor physical and chemical
parameters. Beetle abundance and Carabidae tribe richness were significantly reduced in
dry plots. Community similarity was substantially higher between wet and ambient plots
compared to dry plots due to the substantial reduction of three dominant carabid tribes. Litter
mass increased overall, litter nitrogen decreased, and carbon:nitrogen ratio (C:N) and
total phenolics increased in the dry-plot Oi horizon. Beetle abundance and tribe richness
were positively related to soil moisture, and beetle abundance was negatively related to litter
mass. Microarthropod abundance was highest in the dry treatment. This study provides
evidence that shifting precipitation patterns predicted with climate change could alter important
ground-fauna communities in extensive ecosystems such as temperate forests.
Introduction
Climate is a primary factor shaping the geographic distribution of biota (Coope
and Wilkins 1994). Thus, human-induced alterations in climate elements such as
precipitation patterns may affect the diversity of biota in terrestrial ecosystems on
broad scales. During the past century, the burning of fossil fuels has substantially
increased atmospheric carbon dioxide (IPCC 2007), with consequences that include
rising global mean temperature and changes in precipitation patterns in broad regions
of the planet (IPCC 2007). Changes to the global hydrologic cycle have the
potential to affect plant productivity, biogeochemical cycling, and water resource
availability in many ecosystems, including forests (Hanson and O’Hara 2003).
Specific effects on biodiversity remain largely uncertain and should be explored as
part of comprehensive efforts to predict impacts of changing precipitation patterns
on terrestrial ecosystems (Weltzin et al. 2003).
Potential changes in community composition after long-term exposure to precipitation
alterations may be an especially relevant aspect of diversity to examine
1Department of Biology, PO Box 32027, Appalachian State University, Boone, NC 28608.
2Current address - Department of Sciences, Health and Physical Education, Randolph Community
College, 629 Industrial Park Avenue, Asheboro, NC 27205. 3Current address - Department
of Zoology, 212 Pearson Hall, Miami University, Oxford, OH 45056. 4Oak Ridge
National Laboratory, 1 Bethel Valley Road, Oak Ridge, TN, 37831. *Corresponding author
- willmsrs@appstate.edu.
Manuscript Editor: Wade Worthen
Southeastern Naturalist
139
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
in relation to climate. Walther (2007) argued that species’ reactions to factors such
as drought are part of a complex cascade of reciprocal responses and feedback processes
that could influence other biota and ecosystem functions. Furthermore, some
animal taxa have been recognized as bioindicators and may be useful in studies to
monitor and detect changes in the environment (Bohac 1999, Rainio and Niemela
2003). For example, studies indicate that the length of drought in forests affects the
recovery of key soil arthropods (Lindberg and Bengston 2006), and climate changeinduced
drought affects insect outbreak species (Negrón et al. 2009) and possibly
invasive insect species dynamics (see Dale et al. 2001). Precipitation reduction in
forests can alter the diversity of top invertebrate predators such as spiders (Cramer
2003), which suggests that a close examination of trophic responses to environmental
change is warranted. Understanding precipitation alteration in ecosystems will
clearly benefit from more studies focusing on changes in community composition
and biodiversity.
Ground beetles (Order Coleoptera:Family Carabidae) are an ideal taxon to use
in studies that examine the effects of precipitation change because they have a
high species diversity and are relatively easy to identify taxonomically, are ubiquitous,
easy to sample and sensitive to minor habitat modifications (Desender
1996, Niemela et al. 1996, Rykken et al. 1997). Ground beetles constitute a substantial
fraction of the ground-dwelling fauna in temperate forest (Magura 2002).
Moreover, these beetles are known to quickly colonize areas where suitable habitat
becomes available (Elias 1991), and sensitive species are likely to respond to
habitat changes resulting from precipitation alteration by dispersing to favorable
environments. Carabidae are affected by both moisture availability (Antvogel and
Bonn 2001, Maudsley et al. 2002, Rykken et al. 1997, Yi and Moldenke 2005)
and the physical environment (Magura et al. 2004). Hence, precipitation could
influence ground beetle community structure through ecosystem-level feedbacks
between water availability and biotic processes that alter essential microhabitat
for invertebrate inhabitants of the forest floor (Chikoski et al. 2006, Johnson et
al. 2002, Taylor and Wolters 2005). Examples of such feedbacks include moisture
effects on decomposition and hence forest-floor structure, and moisture effects on
microinvertebrates that can influence decomposition and are also prey species for
macroinvertebrates (Johnston 2000).
The soil environment varies dramatically across the landscapes of the southern
Appalachians, and soil properties that have developed in relation to topographic
position are likely to influence ground-dwelling faunal communities. Moving from
upper to lower slope positions, moisture and nutrient availability generally increase
as soils become more finely textured and drainage is slower (Boerner 2006, Day
and Monk 1974). As a consequence, the response of soil moisture to periods of
drought in southern Appalachian forests depends strongly on topographic features
of the landscape (Yeakley et al. 1998). These dynamics illustrate the importance
of landscape position to soil processes driven by moisture, and these processes can
influence landscape patterns of other characteristics important to ground fauna on
small scales, such as soil carbon (Bolstad and Vose 2001).
Southeastern Naturalist
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
140
This investigation was part of a larger experiment that examined forest responses
to precipitation change over 13 years and provided a unique opportunity to
investigate precipitation-induced effects on important ground fauna in an intact forest
subjected to long-term precipitation alteration. Some researchers have predicted
that the eastern US will experience overall moderate increases in precipitation as
the planet warms (IPCC 2007). The data generated from this experiment shed light
on the likely effects of two extremes in patterns of precipitation—increased precipitation
and drought—on an extensive forest ecosystem. To better understand the
interrelationships between climate alteration, habitat, and ground-dwelling arthropod
community parameters, we sampled arthropods over a single growing season.
The study design allowed us to quantify effects of precipitation manipulation on
communities of Coleoptera (beetles), with a special emphasis on ground beetles
(Family Carabidae). We addressed several questions in our study: (1) does altered
precipitation influence the most prevalent beetle taxa and ground beetle communities
in temperate forests of the Southern Appalachians? (2) are the responses of
ground beetles to precipitation related to chemical or physical changes in the forest
floor? and (3) do ground beetle responses relate to changes in potential prey as influenced
by treatments? We hypothesized that precipitation alteration would result
in community-level effects on dominant taxa such as Carabidae, as groups (e.g.,
tribes) within this large beetle family would shift to preferred habitats. We further
hypothesized that changes in structural and chemical characteristics of the litter due
to treatment would cause insect responses.
Field Site Description
This study was part of a large-scale experiment initiated in 1993 to examine the
long-term effects of precipitation alteration on ecosystem processes in an intact
temperate forest. The throughfall displacement experiment (TDE) was located on a
south-facing slope in Walker Branch Watershed, part of the US Department of Energy’s
(DOE) National Environmental Research Park near Oak Ridge, TN. Walker
Branch is a temperate-zone forest with a mean annual precipitation of 1358 mm
and an average temperature of 14.2 °C (Hanson et al. 2003). The area was chosen
because of its uniform slope, consistent soils, and reasonably uniform distribution
of vegetation. Quercus alba L. (White Oak), Quercus prinus Willd. (Chestnut Oak),
and Acer rubrum L. (Red Maple) were the dominant tree species in the TDE (Hanson
et al 2003).
The overall experimental design at the TDE is described by Hanson and Wullschleger
(2003). Briefly, the site consisted of 3 adjacent 80-m x 80-m treatment areas
(dry, ambient, and wet) positioned at the upper divide of the watershed to avoid
lateral flow into the site from an upslope position. An elevation map developed for
the site shows gradual relief from the upslope to the downslope position, with an
overall change in elevation of 21 m from top to bottom, on average (Hanson et al.
2003). Researchers manipulated hydrologic inputs with a network of 2000 sub-canopy
troughs (0.3 m x 5 m) that diverted an estimated 33% of available precipitation
from the dry to the wet treatment. The ambient treatment, with no precipitation
Southeastern Naturalist
141
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
alteration, served as a control. Each treatment area was divided into one hundred
8-m x8-m plots. Soil-moisture measurements began one year pre-treatment. During
the seven years of precipitation manipulation, researchers observed significant
differences between treatments; wet plots had higher soil moisture compared to
ambient and dry plots except during the dormant season and under extreme drought
(Hanson et al. 2003). These data clearly demonstrate a developing difference in soil
moisture levels between dry plots and those receiving normal or augmented water
as a growing season progresses.
Methods
For this experiment, we randomly selected 10 of the 100 plots in each treatment
area for sampling arthropods and abiotic parameters. Plots used in the study were
free of towers, cages, or other equipment that could alter forest-floor conditions.
We maintained a large buffer between treatments by only interior plots within treatment.
Of the 10 plots in each treatment, we sampled 5 plots in upper slope positions
and 5 plots at lower slope positions to account for any differences caused by
landscape position.
Biotic measurements
In 2005, we used to pitfall traps to collect ground-dwelling arthropods during 3
sampling periods. Traps consisted of 250-ml Nalgene bottles with a 10-cm diameter
funnel inserted at the top and situated so that the funnel mouth was flush with the soil
surface; each bottle contained 200 ml of a 50% ethanol solution for preservation of
specimens. Although this method of collecting ground-dwelling beetles likely measures
activity rather than absolute numbers or density, we believe that differences
in abundance and diversity were still reflected in our captures. To simplify the data
presentation, we use the term abundance, rather than terms such as activity density,
etc., with respect to beetles. We initiated arthropod collections on 11 May, 16 July,
and 21 September 2005, and traps remained open for 5 days and nights for each
collection. We placed 5 traps at 1-m intervals along a linear transect in the center
of each plot running across the slope. At the conclusion of each sample period, we
combined the contents of the 5 traps within a plot into one composite sample, for a
total of 10 samples per treatment per sample date. We sorted contents of traps into
broad taxonomic categories: Aranae (Spiders), Opiliones (Harvestmen), Orthoptera
(Grasshoppers), Formicidae (Ants), Coleoptera (Beetles), Chilipoda/Diplopoda
(Millipedes/Centipedes), and other, which consisted of the Orders Collembola
and Hymenoptera. We classified beetles to the family level, and further identified
ground beetles (Family Carabidae) to tribe following Arnett and Thomas (2001).
Although multiple invertebrate taxa that serve as bioindicators may be necessary
to fully assess environmental change (see Riggins et al. 2009), for the purpose of
this study, we focused our analysis on Coleoptera because of their substantially
greater overall abundance compared to other arthropods and because the sampling
method we used was more suited to this group, reducing the problem of sampling
artifacts. In addition, unlike the situation for other arthropods, previous studies
Southeastern Naturalist
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
142
have described specific habitat requirements for the Family Carabidae and their
data provided a basis for comparison that would not have been available for other
taxa (see Cameron and Leather 2012, Desender 1996, Rykken et al. 1997). Further,
Carabidae response to biotic and abiotic factors at different spatial scales provides
a potentially useful framework for looking at effects of environmental change (Koivula
2011). The use of tribes provided a manageable and ecologically relevant level
of classification based on known characteristics of many groups. Previous work
provides evidence that supra-species groupings are appropriate when univariate
analyses are used to detect large habitat-fragmentation and landscape-level effects,
but less robust on finer scales and with multivariate analyses (see Grimbacher et al.
2007). Additionally, taxonomy at a level such as Family can provide indications of
environmental change in invertebrates inhabiting soil (Riggins et al. 2009), though
groups below this level with considerable variation in trophic response could prove
problematic. We feel that our analysis and hypothesis regarding habitat preferences
supports the use of supra-species groupings.
We calculated the community level parameters of richness and evenness using
both dominant beetle families (see below for a more detailed description) and
ground beetle tribes. Here. richness (R) is the number of families or tribes contained
in a sample. We calculated evenness as E = H'/logeS, where H' is the Shannon
diversity index and S is the number of families or tribes within the sample (formula
following Magurran 2004). Finally, for Carabidae, we calculated the Sørensen
similarity index to compare tribe similarity: S = A/ (B + C) x 200, where letters A,
B and C represent the minimum number of individuals in tribes within treatments
(formula following van Tongeren 1995).
We sampled microarthropods using 5-cm-diameter x 4-cm-depth cores at the
same time as pitfall trap collections. The 4-cm depth included Oe and Oa horizons
and a small amount of mineral soil. We collected 5 cores per plot and combined
them into one sample per plot (total 30 samples/collection date). We kept samples
at 4 °C for no more than 24 h until we extracted the organisms using the methods
of Crossley and Blair (1991) with 10-cm-diameter x 21.5-cm-high PVC extractor
tubes with a mesh screen on the bottom. We fitted each extractor with a 5-watt bulb
light and we controlled intensity with a rheostat. We continued microarthropod
extraction for 7 days, increasing the intensity of light (i.e., heat) with each consecutive
day. We quantified organisms in the Order Acari (mites) and expressed the
counts as number per m2.
Abiotic measurements
We quantified the mass of the O horizons in summer (June 21) and fall (September
22), 2005. We refer to the sum of the 3 O horizons (Oi, Oe and Oa) as
forest floor, and the un-decomposed Oi layer as litter. We sampled the Oi and
Oe horizons by cutting around a 15-cm x 15-cm wooden square and excavating
to the bottom of each horizon. We collected the Oa horizon in cores (5-cmdiam
to the surface of the mineral soil) removed from the center of the excavated
squares. In summer, we collected 4 samples from each plot for a total of 40 samples
in each horizon/treatment. We took our samples from the corner of the plot to
Southeastern Naturalist
143
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
minimize disturbance to ground-dwelling fauna. In the fall, we collected only 2
samples from the Oi and Oe horizon (20 samples/treatment), and we did not sample
the Oa horizon. For our data analysis, we used soil moisture (expressed as %
v/v) values TDE data archive (http://tde.ornl.gov/tdedata.html). We averaged biweekly
measurements throughout 2005 using a time domain reflectometer (TDR)
technique at the 0–35 cm soil layer for each treatment (Hanson et al. 2003). Average
soil moisture was 21.2 (wet), 20.3 (ambient) and 17.7 (dry). We recorded
surface temperature using iButton continuous data loggers (Maxim Integrated
Products, San Jose, CA) positioned in the center of each plot directly beneath
the Oi layer of the organic horizon. We programmed data loggers to record
temperature at 4-h intervals from 9 May–28 September 2005. We downloaded
temperature data using iButton TMEX Application software and calculated weekly
minimum, maximum, and mean temperatures
We measured total carbon (C), total nitrogen (N), and carbon:nitrogen ratio
(C:N) for the Oi, Oe, and Oa horizons using samples collected in May 2005. We
pooled dried samples within each plot and ground each one in a coffee mill and
then in a ball mill until the material had a talcum powder consistency. We sent 1
sample per plot (total 10 per treatment) from each of the Oi, Oe, and Oa horizons to
the University of Georgia Institute of Ecology Stable Isotope Laboratory (Athens,
GA) for aqnalysis of total C and N using the Micro-Dumas combustion technique.
We analyzed total phenolic content in the same Oi, Oe and Oa samples following
the Folin-Ciocalteu (FC) reagent technique of Singleton and Rossi (1965) and expressed
the values as percent tannic acid equivalents (%TAE).
Statistical analyses
We tested effects of treatment on dominant beetle family and Carabid tribe community
parameters (i.e., abundance, richness, and evenness) and mite abundance
using one-way ANOVA (Proc GLM, SAS 9.1, SAS Institute, Cary, NC) with all
samples combined (hereafter, the cumulative dataset). We log transformed the data
to achieve normality. We present untransformed data where appropriate. We used
Tukey’s honestly significant difference (HSD) test for selected insect communityparameter
post hoc pair-wise comparisons of treatment means. We used Proc GLM
to test effects of treatment on average forest-floor mass and moisture; N (%); C:N
ratio; total phenolic content from the Oi, Oe, and Oa horizons; and mean weekly
temperature minima and maxima. For all analyses, replication was at the level of
plot, and significance assigned at P ≤ 0.05. We report results where 0.10 ≥ P ≥ 0.05
as marginally significant. We fully acknowledge the un-replicated design of the
larger TDE experiment and suggest that this approach is reasonable when such
costly experimental field designs are undertaken (see Eberhardt and Thomas 1991)
and where sufficient pre-treatment data exist to demonstrate that any observed
treatment effects are not due to variations across sites before the experiment began.
We examined variables potentially important to ground fauna, including dominant
trees, soils, microclimate, slope, and patterns of soil moisture, prior to setting up the
experiment (Hanson et al 2003). Forest-stand and understory species composition
Southeastern Naturalist
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
144
and basal area were not different between sites (Hanson et al. 2001) and a comparison
of archived data (http:tde.ornl.gov) demonstrates strong similarity in soil
characteristics. Along with numerous other authors, we conclude that sufficient
similarity existed between sites prior to the manipulation treatments to allow for
treatment comparisons. Based primarily on an initial correlation analysis (Proc
CORR; SAS) that examined relationships between fauna and physical features
of the forest floor, we used linear regression (Proc REG) to analyze relationships
between beetle and carabid community parameters and principal abiotic variables.
We used the same approach for cumulative mite abundance. We also used partial
least squares regression (PSLR), where all soil and litter variables (moisture, N,
CN, %TAE, etc.) are simultaneously loaded into this permutation procedure to
generate predicted values for carabid community parameters. A linear regression
of actual versus predicted values provided a measure of how the abiotic variables
related to beetle abundance, richness, and evenness. This procedure is gaining wide
aceptance in the ecological literature and is seen as an appropriate alternative to
more classical regression analyses (Carrascal et al. 2009).
Results
Forest-floor biota
We collected a total of 3244 beetles from 29 families during 2005 at the TDE:
1296 from the wet treatment, 1045 from the ambient treatment, and 903 from
the dry treatment. Seven families of beetles comprised greater than 96% of the
abundance. Twenty families had 9 or fewer individuals in traps in 2005, and in
our estimation, these may have represented random captures. For this reason, our
initial analysis focused on the 7 dominant beetle families; Carabidae, Curculionidae,
Nitulidae, Staphylinidae, Scolytidae, Scarabaeidae, and Chrysomelidae, for a
total of 3120 beetles. There was no treatment bias towards the number of beetles
in other families excluded from the analysis: wet = 44, ambient = 52 and dry = 55.
Two families—Carabidae and Curculionidae—comprised approximately 57% of
all dominant beetles collected (Table 1).
There was considerable variation in percent abundance between precipitation
treatments in the 7 dominant beetle families, with no treatment difference observed
(Tables 1, 2). Family-level richness was marginally affected by treatment (Table 2),
and richness was lowest in plots where water was intercepted (mean ± SE; wet: 6.5
± 0.2; ambient: 6.8 0.1; dry: 6.2 ± 0.2). The higher percentage of the (Carabidae
and Curculionidae) in the wet plots compared to ambient or dry plots (Table 1)
likely contributed to a significant effect on evenness (Table 2), which was lowest
in plots that received additional water (wet: 0.792 0.03; ambient: 0.837 ± 0.03;
dry: 0.918 ± 0.1).
We collected a total of 1070 Carabidae from 9 tribes (Table 1). Overall, carabid
beetles were more prevalent in wet or ambient plots than in dry plots. The tribes
Harpalini, Callistini, and Pterostichini made up approximately 82% of all Carabidae
in our samples (Table 1). As a percentage of total abundance, beetles in the tribe
Southeastern Naturalist
145
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
Harpalini were much more prevalent in wet plots compared to dry plots; in the tribe
Pterostichini ,the difference between wet and dry plots was less pronounced. The
abundance-based Sørensen similarity index indicated a substantial effect of drought
treatment on community composition at the tribe level. Communities in the wet
and ambient treatments were 90% similar, those in the dry and wet treatments were
56% similar and those in the dry and ambient treatments were 69% similar. These
community differences appeared to be due largely to treatment effects on the 3
dominant tribes. Responses of some other tribes to treatment probably contributed
to differences in community similarity, though we observed lower abundances and
less consistent differences between treatments (Table 1).
Precipitation treatment clearly altered total beetle abundance and tribe richness
within the Carabidae (Table 2). Beetle abundance was 61% lower in the dry treatment
compared to the wet and ambient treatments (P = 0.05, Table 1, Fig. 1A), and
tribe richness was lower in the dry than wet and ambient treatments (Table 1, Fig.
1B). Tribe-level evenness was not affected by treatment (Table 2, Fig. 1C).
Table 1. Percent average abundance by treatment for the seven dominant beetle families and Tribes in
Family Carabidae. Numbers calculated from cumulative abundance data.
n Wet Ambient Dry
Family
Carabidae 1070 45.7 37.2 17.1
Curculionidae 716 53.8 20.8 25.4
Staphylinidae 465 30.8 29.2 40.0
Nitulidae 394 32.2 33.5 34.3
Scolytidae 279 26.2 38.0 35.8
Scarabaeidae 132 23.5 32.6 43.9
Chrysomelidae 64 25.0 60.9 14.1
All families (N) 3120 1061 1121 938
Carabidae Tribes
Harpalini 409 55.3 36.9 7.8
Pterostichini 362 39.2 37.0 23.8
Callistini 75 45.3 44.0 10.7
Licinini 61 52.5 29.5 18.0
Galeritini 51 27.5 41.2 31.4
Cychrini 48 33.3 35.4 31.3
Cicindelid 26 23.1 46.2 30.8
Notiophilinini 22 50.0 27.3 22.7
Scaratini 16 68.8 12.5 18.8
All tribes (N) 1070 492 394 184
Table 2. ANOVA results for the effects of treatment on dominant beetles and Carabidae tribe community
parameters. df = 2, 27; n = 30; * = P ≤ 0.05.
Abundance Richness Evenness
F P F P F P
All beetles 2.07 0.145 3.16 0.060 6.52 0.005*
Carabidae tribes 13.1 0.0001* 4.03 0.029* 1.24 0.306
Southeastern Naturalist
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
146
Mite abundance (mean ± SE) was affected by treatment (df = 2, 24; P = 0.024),
with more mites found in the dry (228 ± 28) than wet (156 ± 35) or ambient (134
± 13) plots.
Figure 1. Carabidae community parameters
(mean + SE) by treatment
for cumulative (A) abundance, (B)
tribe richness, and (C) tribe evenness
with slope combined. * = significant
treatment difference (P ≤ 0.05),
Tukey’s HSD test.
Southeastern Naturalist
147
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
Forest-floor characteristics
Intercepting water from the canopy resulted in total forest-floor mass in the dry
treatment approximately 2 times greater than in the ambient, and 1.5 times greater
than in the wet treatment (Table 3). All layers, with the exception of the Oe, were
significantly affected by treatment, and forest-floor mass was highest in the dry
plots (Table 3). Weekly average forest floor minimum, maximum, and mean temperatures
were not affected by treatment or slope during any sampling period (data
not shown).
Effects of treatment on forest-floor litter chemistry—%N, C:N ratio and total
phenolics—was dependent on the O horizon. We observed significant effects of
treatment only in the Oi layer for nutrients, where %N was lower and C:N ratio
higher in dry plots compared to the ambient and wet plots (Table 3). Total phenolics
were greater in the Oi horizon in the dry plots, and we observed a marginally
significant reduction in this measure in the Oa horizon (i.e., h umic layer).
Regression analyses
The principal predictors of beetle community parameters were soil moisture and
mass of the forest floor. Carabid abundance (P = 0.007, R2 = 0.34) and tribe richness
(P = 0.05, R2 = 0.13) were positively and significantly related to soil moisture (Fig.
2A, B). Tribe evenness was marginally negatively related to soil moisture (P = 0.08,
R2 = 0.11; Fig. 2C). Carabid abundance was marginally negatively related to forest-
Table 3. Mean, standard error (SE), F Ratio, P valueA, and N (Proc GLM) for organic horizon mass,
%N and C:N ratio and phenolics (%TAE).* = P ≤ 0.05
Wet Ambient Dry
Mean SE Mean SE Mean SE F P
Mass (g m-2)
Oi 354 17 296 12 399 25 7.63 0.002*
Oe 495 28 537 21 542 31 1.06 0.361
Oa 2417 264 1326 265 3231 322 8.11 0.002*
Total 2782 496 2026 276 4173 342 6.29 0.006*
N (%)
Oi 1.31 0.06 1.30 0.03 1.11 0.07 4.62 0.019*
Oe 1.56 0.08 1.40 0.06 1.45 0.04 1.85 0.117
Oa 1.12 0.07 1.08 0.06 1.26 0.06 1.54 0.237
C:N ratio (mg mg-1)
Oi 37.2 1.6 37.1 1.2 44.2 2.7 4.02 0.031*
Oe 30.5 1.3 31.0 1.7 31.8 1.5 0.20 0.827
Oa 27.2 1.5 32.1 5.2 26.1 1.1 1.06 0.366
Phenolics (%TAE)
Oi 10.9 0.6 11.0 0.6 13.3 0.9 3.71 0.038*
Oe 6.51 0.4 6.36 0.6 6.60 0.4 0.06 0.942
Oa 14.0 1.7 8.28 1.2 12.6 1.9 2.76 0.087
AOi, Oe, mass, total mass, Oe N% and C:N, Oi and Oe %TAE: d. f. = 2, 27, n = 30. Oi N% and C:N:
df = 2, 25; n = 28. Oa mass: df = 2, 24; n = 27. Oa N%, C:N and %TAE: df = 2, 21; n = 24.
Southeastern Naturalist
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
148
floor mass (P = 0.08, R2 = 0.11, Fig. 3A), and tribe richness was unrelated to mass
(P = 0.217, R2 = 0.05, Fig. 3B). Tribe evenness marginally increased (P = 0.07, R2 =
0.11, Fig. 3C) with higher forest-floor mass. Partial least squares regression found
Figure 2. Relationship between soil moisture
(% volume/volume ) and cumulative
(A) Carabidae abundance (P = 0.007, R2 =
0.34), (B) tribe richness (P = 0.05, R2 =
0.13), and (C) tribe evenness (P = 0.08,
R2 = 0.11).
Figure 3. Relationship between forest floor
mass (Oi + Oe + Oa) and cumulative (A)
Carabidae abundance (P = 0.08, R2 = 0.11),
(B) tribe richness (P = 0.217, R2 = 0.05), and
tribe evenness (P = 0.07, R2 = 0.11).
Southeastern Naturalist
149
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
relationships between forest-floor structural and chemical parameters with carabid
abundance (P = 0.001, R2 = 0.52), richness (P = 0.001, R2 = 0.40), and evenness
(P = 0.030, R2 = 0.28) (Fig. 4).
Figure 4. Relationship between all soil
and litter abiotic variables on actual versus
predicted (A) Carabidae abundance
(P = 0.001, R2 = 0.53), (B) tribe richness
(P = 0.001, R2 = 0.40), and tribe evenness
(P = 0.003, R2 = 0.28) using partial least
squares regression.
Southeastern Naturalist
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
150
Mite abundance was related to key physical characteristics of the forest floor.
Abundance decreased with increasing soil moisture (P = 0.006, R2 = 0.24) and
was positively related to forest-floor mass (P = 0.007, R2 = 0.23). None of the
carabid community parameters was significantly related to mite abundance (data
not shown).
Discussion
Manipulating the amount of natural precipitation that reached the forest floor affected
the distribution of dominant beetle families, with resultant changes in community
measures in a prevalent beetle family. Intercepting precipitation reduced
the abundance of beetles and decreased the richness of tribes within the Carabidae,
which comprised a substantial component of the macroarthropod community.
Community-level effects of reduced precipitation were clearly evident in the higher
similarity between beetle communities of wet and ambient treatments compared
to the dry treatment. Our data provides evidence that key community parameters
within the Family Carabidae could change in response to reduced precipitation, and
that colonization of preferred habitats created by climate change is likely for certain
tribes in this large family of beetles. Important forest-floor characteristics such as
soil moisture, litter mass, and chemistry changed due to alterations in precipitation,
which in turn may have affected insect responses. These results contribute to
a better understanding of the potential effects of altered precipitation in temperate
forests in at least 2 important ways. First, the large scale (1.92 ha) of the precipitation
experiment allowed us to examine effects on the habitat and arthropods across
an extensive landscape. Second, the duration (13 years) of the larger experiment enabled
us to study the long-term impact of altered precipitation in forests. Although
we sampled in only a single year, we conclude that observed changes in the physical
characteristics of the forest floor that affected beetles had accumulated over the
many years of the larger TDE experiment, resulting in the observed effects on the
beetle community over a much longer time than a brief sampling period.
Our results suggest that ground beetle communities responded to precipitation
changes relative to the amount of water that reached the forest floor. This finding
has implications for regions that may experience future drought due to lower
precipitation, or conversely, areas that become wetter due to shifting precipitation
patterns. Two prevalent taxa of Family Carabidae responded to precipitation alterations,
though in slightly different ways. The dominant Tribe Harpalini clearly
shifted to wetter plots, whereas another common tribe, Pterostichini, was somewhat
less responsive to dry conditions even though their abundance was higher in the
more moist plots (Table 1). Combined with the data on other tribes and dominant
beetle families, these findings suggest that important ground-dwelling arthropods
will respond to precipitation changes in different ways. In addition to affects on
species abundance, our data shows that tribe-level richness is negatively affected
by dry conditions, and that a shift in preference for wetter habitat results in community-
level shifts in the distribution of dominant taxa. In the only other study on
the TDE experimental site that examined ground-dwelling fauna (spiders), Cramer
Southeastern Naturalist
151
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
(2003) found evidence of preferences by certain spider species for either the wet
or dry plots. Overall, our observations on abundance, richness, and evenness of the
carabid beetle community strongly supported our hypothesis that long-term alterations
of forest-floor physical characteristics influence community parameters in an
intact deciduous forest.
Water availability in forest-floor soils could also have important effects on carabid
communities and contribute to the lower abundance and tribe-level richness
that we observed in plots where water was intercepted. Soil water-content was
a key predictor of the carabid community in our study, which is consistent with
other similar studies (Luff et al. 1989, Maudsley et al. 2002, Rykken et al. 1997).
Ground beetle abundance and tribe-level richness were positively related to soil
moisture content, with each measure generally higher in the wet than dry treatment.
The specific moisture requirements of certain species are known to influence
carabid community dynamics (Maudsley et al. 2002, Rykken et al. 1997). Other
studies determined that dry conditions reduce overall beetle abundance and specific
Carabidae taxa (Kiovula et al. 1999, Yi and Moldenke 2005). The tribe Harpalini
was more abundant in the wet than dry treatment, consistent with findings that the
distribution of a principal genus in this tribe was constrained by moisture deficits
(Noonan 1990). Our data shows that increased precipitation in a deciduous forest
could benefit this moisture-preferring taxon.
It is also likely that the indirect effects and feedbacks of soil water availability
are contributing to the community patterns that we found, including the preference
by the tribe Harpalini for the wet treatment. Though only marginally significant,
ground beetle abundance was inversely related to the mass of the forest floor, suggesting
that precipitation influences ground beetle communities through effects on
the forest floor. Our results contrast with previous litter- or resource-addition studies,
where arthropod communities responded positively to increased litter depth and
associated changes in architecture (Bultman and Uetz 1984, Halaj and Wise 2002,
Kiovula et al. 1999).
In our study, reduced water availability likely contributed to the greater total
forest-floor mass found in the dry treatment compared to the ambient and wet treatments.
Consistent treatment effects on soil water content have been demonstrated
in long-term data sets at the TDE (Hanson and Wullschleger 2003). This result has
implications for important forest-floor processes, because the frequency and intensity
of drying-rewetting cycles are known to affect microorganisms responsible for
the vast majority of decomposition in natural systems (Fierer at al. 2003, Schimel
et al. 1999). The finding of lower %N and higher C:N ratio in the litter layer of the
dry treatment plots is consistent with reduced decay rates, because N is generally
enriched relative to C as decomposition progresses (Taylor et al. 1989). Diverting
water may also have influenced decomposition in the dry plots by increasing
carbon-based phenolics measured as tannic acid equivalents, thus potentially affecting
decay processes (Gallardo and Merino 1993, Taylor et al. 1989). We found
higher %TAE in the un-decomposed Oi (i.e., litter) layer. It seems likely, based
on the chemical analyses, that removing natural precipitation results not only
in litter accumulation but also in the production of a more slowly decomposing,
Southeastern Naturalist
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
152
lower-quality litter, each of which could negatively affect important arthropod
taxa. When all abiotic parameters were considered, our study showed that changes
in precipitation affected physical and chemical characteristics, which, in turn, may
have affected the ground beetle community.
One indirect way that precipitation can alter forest beetle communities is by affecting
trophic-level interactions. Ground beetle responses to precipitation in this
study were not directly related to changes in microarthropods, which are potentially
important prey species in the forest floor. In contrast to ground beetles, the abundance
of mites increased in the dry treatment and declined relative to soil moisture across
treatments. One possible explanation is that the larger number of beetles in wet plots
reduced the mite abundance by predation. It is also possible that litter accumulation
due to drought creates a more favorable physical environment for mites (Hansen
2000). In addition to this effect on the physical environment, drier conditions could
increase mite abundance if it simultaneously reduced predatory beetle abundance
through negative effects of dry soil on beetle larvae (see Loreau 1987).
Conclusions
This study found that precipitation manipulation over 13 years in an intact forest
altered dominant beetle families and in particular, ground beetle communities, possibly
through effects on forest-floor mass, soil moisture, and soil chemistry, which
altered the structure and quality of the beetle habitat. This work points to the need
to further pursue specific effects on the distribution of dominant species in Carabidae
relative to environmental change. Our study provides insight into the influence
of precipitation alteration on abiotic and biotic components of the forest floor in
temperate hardwood forest landscapes, and our reslts increase our understanding of
terrestrial ecosystem responses to future climate change.
Acknowledgments
Special thanks to go to Shawn Villalpando (ASU) for his field assistance and Don Todd
(ORNL) for his help at the TDE field laboratory site. Support for the TDE experiment was
obtained from the US Department of Energy (DOE), Office of Science, Biological and
Environmental Research (BER) program as part of the Program for Ecosystem Research
(PER). Oak Ridge National Laboratory is managed by UT-Battelle, LLC, for the DOE under
contract DE-AC05-00OR22725.
Literature Cited
Antvogel, H., and A. Bonn. 2001. Environmental parameters and microspatial distribution
of insects: A case study of carabids in an alluvial forests. Ecography 24:470–482.
Arnett, H.G., Jr., and M.C. Thomas. 2001. American Beetles: Archostemata, Myxophaga,
Adephaga, Polyphaga: Staphyliniformia. CRC Press, Boca Raton, FL. 443 pp.
Boerner, R.E.J. 2006. Unraveling the Gordian knot: Interactions among vegetation, topography,
and soil processes in the central and southern Appalachians. Journal of the Torrey
Botanical Society 133:321–361.
Bohac, J. 1999. Staphylinid beetles as bioindicators. Agriculture Ecosystems and Environment
74:357–372.
Southeastern Naturalist
153
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
Bolstad, P.V., and J.M. Vose. 2001. The effects of terrain position and elevation on soil C
in the southern Appalachians. Pp. 45–52, In R. Lall, J.M. Kimble, R.F Follett, and B.A.
Stewart (Eds.). Advances in Soil Science: Assessment Methods for Soil Carbon. Lewis
Publishers (CRC Press), Boca Raton, FL. 654 pp.
Bultman, T.L., and G.W. Uetz. 1984. Effect of structure and nutritional quality of litter on
the abundance of litter-dwelling arthropods. American Midland Naturalist 111:165–172.
Cameron, H., and S.R. Leather. 2012. How good are carabid beetles (Coleoptera, Carabidae)
as indicators of invertebrate abundance and order richness? Biodiversity and
Conservation 21:763–779.
Carrascal, L.M., I. Galván, and O. Gordo. 2009. Partial least squares regression as an alternative
to current regression methods used in ecology. Oikos 118:681–690.
Coope, R.G., and A.S. Wilkins. 1994. The response of insect faunas to glacial-interglacial
climatic fluctuations. Philosophical Transactions of the Royal Society of London Series
B 344:19–26.
Chikoski, J.M., S.H. Ferguson, and L. Meyer. 2006. Effects of water addition on soil arthropods
and soil characteristics in a precipitation-limited environment. Acta Oecologia
30:203–211.
Cramer, K.L. 2003. The influence of precipitation change on spiders as top predators in the
detrital community. Pp. 347–359, In P.J. Hanson and S.D. Wullschleger (Eds.). North
American Temperate Deciduous Forest Responses To Changing Precipitation Regimes.
Springer, New York, NY. 472 pp.
Crossley, D.A., and J.M. Blair. 1991. A high-efficiency, low-technology Tullgren-type extractor
for soil microarthropods. Agriculture Ecosystems and Environment 34:87–192.
Dale, V.H., L.A. Joyce, S. McNulty, R.P. Neilson, M.P. Ayers, M.D. Flannigan, P.J. Hanson,
L.C. Irland, A.E. Lugo, C.J. Peterson, D. Simberloff, F.J. Swanson, J. Stocks, and B.M.
Wolton. 2001. Climate change and forest disturbance. Bioscience 51:723–734.
Day, F.P., and C.D. Monk. 1974. Vegetation patterns on a southern Appalachian watershed.
Ecology 55:1064–1074.
Desender, K.R. 1996. Diversity and dynamics of coastal dune carabids. Annales Zoologici
Fennici 33:65–75.
Eberhardt, L.L., and J.M. Thomas. 1991. Designing environmental field studies. Ecological
Monographs 61:53–73.
Elias, S.A. 1991. Insects and climate change. Bioscience 41:552–559.
Fierer, N., J.P. Schimel, and P.A. Holden. 2003. Influence of drying-rewetting frequency on
soil bacterial community structure. Microbial Ecology 45:63–71.
Gallardo, A., and J. Merino. 1993. Leaf decomposition in two Mediterranean ecosystems of
southwest Spain: Influence of substrate quality. Ecology 74:152–161.
Grimbacher, P.S., C.P Catterall, and R.L. Kitching. 2007. Detecting the effects of environmental
change above the species level with beetles in a fragmented tropical rainforest
landscape. Ecological Entomology 33:66–79.
Halaj, J., and D.H. Wise. 2002. Impact of a detrital subsidy on trophic cascades in a terrestrial
grazing food web. Ecology 83:3141–3151.
Hanson, P.J., and F.M. O’Hara. 2003. Introduction. Pp 8–31, In P.J. Hanson and S.D. Wullschleger
(Eds.). North American Temperate Deciduous Forest Responses To Changing
Precipitation Regimes. Springer, New York, N.Y. 472 pp.
Hanson, P.J., and S.D. Wullschleger. 2003. North American Temperate Deciduous Forest
Responses To Changing Precipitation Regimes. Springer, New York, N.Y. 472 pp.
Southeastern Naturalist
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
154
Hanson, P.J., D.E. Todd, and J.S. Amthor. 2001. A six-year study of sapling and large-tree
growth and mortality responses to natural and induced variability in precipitation and
throughfall. Tree Physiology 21:345–358.
Hanson, P.J., M.A. Huston, and D.E. Todd. 2003. Walker Branch throughfall displacement
experiment. Pp. 3–7, In P.J. Hanson and S.D. Wullschleger (Eds). North American Temperate
Deciduous Forest Responses To Changing Precipitation Regimes. Springer, New
York, N.Y. 472 pp.
Hansen, R.A. 2000. Effects of habitat complexity and composition on a diverse litter microarthropod
assemblage. Ecology 81:1120–1132.
Inter-governmental Panel on Climate Change (IPCC). 2007. Fourth Assessment Report
(AR4). Available online at http://www.ipcc.ch. Accessed July 2010.
Johnson, D.W., P.J. Hanson, and D.E. Todd. 2002. The effects of throughfall manipulation
on soil leaching in a deciduous forest. Journal of Environmental Quality 31:204–216.
Johnston, J.M. 2000. The contribution of microarthropods to aboveground food webs: A
review and model of belowground transfer in a coniferous forest. American Midland
Naturalist 143:226–238.
Koivula, M.J. 2011. Useful model organisms, indicators, or both? Ground beetles (Coleoptera,
Carabidae) reflecting environmental conditions. ZooKeys 10 0:287–317.
Koivula, M., P. Punttila, Y. Haila, and J. Niemela. 1999. Leaf litter and the small-scale
distribution of carabid beetles (Coleoptera, Carabidae) in the boreal forest. Ecography
22:424–435.
Lindberg, N., and J. Bengston. 2006. Recovery of forest soil fauna diversity and composition
after repeated summer droughts. Oikos 114:494–506.
Loreau, M. 1987. Vertical distribution of activity of carabid beetles in a beech forest floor.
Pedobiologia 30:173–178.
Luff, M.D., M.D. Eyre, and S.P. Rushton. 1989. Classification and ordination of habitats of
ground beetles (Coleoptera, Carabidae) in north-east England. Journal of Biogeography
16:121–130.
Magura, T. 2002. Carabids and forest edge: Spatial pattern and edge effect. Forest Ecology
and Management 157:23–37.
Magura, T., B. Tothmeresz, and Z. Elek. 2004. Effects of leaf-litter addition on carabid
beetles in a non-native Norway Spruce plantation. Acta Zoological Academy of Sciences
Hungary 50:9–23.
Magurran, A.E. 2004. Measuring Biological Diversity. Blackwell Publishing, Malden, MA.
256 pp.
Maudsley, M., B. Seeley, and O. Lewis. 2002. Spatial distribution patterns of predatory
arthropods within an English hedgerow in early winter in relation to habitat variables.
Agriculture Ecosystems and Environment 89:77–89.
Negrón, J.F., J.D. McMillin, J.A. Anhold, and D. Coulson. 2009. Bark beetle-caused mortality
in a drought-affected Ponderosa Pine landscape in Arizona, USA. Forest Ecology
and Management 257:1353–1362.
Niemela, J., Y. Haila, and P. Punttila. 1996. The importance of small-scale heterogeneity in
boreal forests: Variation in diversity of forest-floor invertebrates across the succession
gradient. Ecography 19:352–368.
Noonan, G.R. 1990. Biogeographical patterns of North American Harpalus Latreille (Insecta:
Coleoptera: Carabidae). Journal of Biogeography 17:583–614.
Rainio, J., and J. Niemela. 2003. Ground beetles (Coleoptera:Carabidae) as bioindicators.
Biodiversity and Conservation 12:487–506.
Southeastern Naturalist
155
R.S. Williams, B.S. Marbert, M.C. Fisk, and P.J. Hanson
2014 Vol. 13, No. 1
Riggins, J.J., C.A. Davis, and W.W. Hoback. 2009. Biodiversity of belowground invertebrates
as an indicator of wet-meadow restoration success (Platte River, Nebraska).
Restoration Ecology, 17:495–505.
Rykken, J.J., D.E. Capen, and S.P Mahabir. 1997. Ground beetles as indicators of land-type
diversity in the Green Mountains of Vermont. Conservation Biology 11:522–530.
Schimel, J.P., J.M. Gulledge, J.S. Clein-Curley, J.E. Lindstrom, and J.F. Braddock. 1999.
Moisture effects on microbial activity and community structure in decomposing birch
litter in the Alaskan taiga. Soil Biology and Biochemistry 31:831–838.
Singleton, V.L., and J.A. Rossi. 1965. Colorimetry of total phenolics with phosphomolybdic-
phosphotingstic acis reagents. American Journal of Enology and Viticulture
16:144–158.
Taylor, A.R., and V. Wolters. 2005. Responses of oribatid mite communities to summer
drought: The influence of litter type and quality. Soil Biology and Biochemistry
37:2117–2130.
Taylor, B.R., D. Parkinson, and W.J. Parsons. 1989. Nitrogen and lignin content as predictors
of litter decay rates: A microcosm test. Ecology 70:97–104.
van Tongeren, O.F.R. 1995. Cluster analysis. Pp. 174–212, In R.H.G Jongman, C.F.J. ter
Braak, and O.F.R. van Tongeren (Eds.). Data Analysis in Community and Landscape
Ecology. Cambridge University Press, Cambridge, UK. 324 pp.
Walther, G.R. 2007. Tackling ecological complexity in climate impact research. Science
315: 606–607.
Weltzin, J.F., M.E. Loik, S. Schwinning, D.G. Williams, P.A. Fay, B.M. Haddad, J. Harte,
T.E. Huxman, A.K. Knapp, G. Lin, W.T. Pockman, M.R. Shaw, E.E. Small, M.D. Smith,
S.D. Smith, D.T. Tissue, and J.C. Zak. 2003. Assessing the response of terrestrial ecosystems
to potential changes in precipitation. Bioscience 53:941–952.
Yeakley, J.A., W.T. Swank, L.W. Swift, G.M Hornberger, and H.H. Shugart. 1998. Soilmoisture
gradients and controls on a southern Appalachian hillsplope from drought
through recharge. Hydrology and Earth Systems Sciences 2:41–49.
Yi, H., and A. Moldenke. 2005. Response of ground-dwelling arthropods to different thinning
intensities in young Douglas Fir forests of western Oregon. Environmental Entomology
34:1071–1080.