Transcript Document

Ant Community Changes Associated With
Introduced Plant Species.
Biology Department
Georgetown University
Washington, D.C.
20057
Daniel S. Kjar and Edward M. Barrows
Introduction
Materials and Methods
Native Plant Richness
• We utilized satellite imagery and Geographic
Imaging System data (GIS) with ArcView® and
the AlaskaPak to generate 60 random sites within
the DMWP. We used a Global Positioning System
(GPS) to locate these sites (Figure 1).
• We placed a pitfall trap at the center of each 1m2 site (Figure 2). Pitfalls were run for 24 hr
during a warm night at the end of each trapping
month (June, August, October during 2002–2003).
• We recorded all plant species and relative
coverage for each site during early August of 2002
and 2003 (Figure 2). We recorded the dbh and
species of all trees within a 6-m diameter circle of
each site.
• We took one soil core (40 x 80 mm) from each
site at the end of each trapping period. We used a
modified Berlese-Tullgren design to extract
arthropods from the 60 soil cores (Figure 2). Soil
cores were used to determine soil moisture and
soil structure.
• The results presented here are preliminary.
These data are from 40 sites, do not include soil
core samples, and do not include the 2003
trapping year.
50
40
40
30
30
Total Ant Abundance
Total Ant Abundance
+
20
10
0
0.0
.6
0
.8
.1
1.0
.2
.3
.4
Soil Moisture
Figure 4b. Ants and soil moisture.
A model of environmental factors significantly
associated with changes in ant abundance. Arrow
width represents amount of variance explained by the
association. Arrow direction represents hypothesized
direction of relationship.
Table 2. Regression models for ant
community associations. a,b,c
Figure 4a. Percent invasion, tree
richness, and soil moisture account
for nearly half the variance in ant
abundance (Hierarchical Linear
Regression).
Model Summary
Model
1
2
3
R
.451a
.625b
.672c
Adjusted
R Sq uare
.183
.358
.405
R Sq uare
.204
.391
.451
Std. Error of
the Estimate
11.23765
9.96071
9.58545
a. Predictors: (Constant), Soil Moisture
b. Predictors: (Constant), Soil Moisture, Tree Richness,
c. Predictors: (Constant), Soil Moisture, Tree Richness,
Percent Alien Plant Coverage
50
Aphaenogaster
rudis (a common
ant species)
40
A.
.4
10
Ants
+
Figure 1. (A) To locate sites we used a Trimble® GPS unit
(left, Keith Post with GPS). (B) Satellite imagery and
ArcView® with AlaskaPak were used to generate random
field site loactions in the Dyke Marsh Wildlife Preserve low
forest.
.2
20
Percent Alien Plant Coverage
Invasive Plants
B.
30
Table 4. Regression models for A. rudis
associations. a,b,c
20
ANOVAc
R
Adjusted
R Sq uare
.154
.210
R Sq uare
.176
.250
.419a
.500b
Std. Error of
the Estimate
9.27474
8.96340
Model
1
2
a. Predictors: (Constant), Total Plant Richness
10
Table 5. ANOVA table for A. rudis
associations. a,b,c
Model Summaryc
Model
1
2
Sum of
Squares
696.809
3268.791
3965.600
992.923
2972.677
3965.600
Reg ression
Residual
Total
Reg ression
Residual
Total
df
1
38
39
2
37
39
Mean Square
696.809
86.021
F
8.100
Sig .
.007a
496.462
80.343
6.179
.005b
b. Predictors: (Constant), Total Plant Richness, Alien
Plant Coverage (%)
a. Predictors: (Constant), Total Plant Richness
c. Dependent Variable: Aphaenog aster rudis
b. Predictors: (Constant), Total Plant Richness, Alien Plant Coverage (%)
c. Dependent Variable: Aphaenogaster rudis
0
-1
0
1
2
3
4
5
6
7
Tree Richness
40
40
30
30
20
20
Figure 4c. Ants and tree richness.
C.
Table 3. ANOVA table for ant
community associations. a,b,c,d
D.
ANOVAd
Model
1
2
3
Reg ression
Residual
Total
Reg ression
Residual
Total
Reg ression
Residual
Total
Sum of
Squares
1227.575
4798.825
6026.400
2355.416
3670.984
6026.400
2718.692
3307.708
6026.400
df
1
38
39
2
37
39
3
36
39
Mean Square
1227.575
126.285
F
9.721
Sig .
.003a
1177.708
99.216
11.870
.000b
906.231
91.881
9.863
.000c
a. Predictors: (Constant), Soil Moisture (%)
Figure 2. (A) Soil core tool, (B) plant coverage grid, (C)
pitfall trap, (D) modified Berlese-Tullgren funnel
setup.
Preliminary Results
Images from the low forest in Dyke Marsh Wildlife Preserve: (A)
a tree fall area dominated by Porcelainberry, (B, D) areas of
relatively native forest, (C) an Asiatic bittersweet vine strangling
a tree.
Tree
Richness
Soil Moisture
50
• The study plots had 55 plant species of different
abundances, including 10 invasive species. We found that
the average level of coverage by invasive plant species for
the DMWP forest was 45%, ranging from 0-94% at each
site. Lonicera japonica and Celastrus orbiculatus were the
more common alien plants (Table 1).
• Invasive plant coverage is correlated with decreasing
native plant richness (Figure 3, Appendix: Table 1).
• Increasing alien plant coverage is associated with
increasing abundances in the native ant community. Tree
richness and soil moisture are also highly correlated with
ant abundance (ANOVA, P < 0.05, Figures 4a, 4b, 4c,
Tables 2 and 3).
• Aphaenogaster rudis may be a good indicator of
ecological change caused by invasion of alien plants
(ANOVA, P < 0.05, Figures 5a, 5b, Appendix: Table 3).
b. Predictors: (Constant), Soil Moisture (%), Tree Richness
c. Predictors: (Constant), Soil Moisture (%), Tree Richness, Alien Plant Coverage (%)
d. Dependent Variable: Ant Abundance
Soil Moisture
Invasive Plants
12
10
8
Ampelopsis brevipedunculata 9.28%
Celastrus orbiculatus
24.59%
Clemitis terniflora
2.55%
Duchesnea indica
< 0 .01%
Euonymus fortunei
0.93%
Invasive Plant Coverage
Hedera helix
3.25%
Ligustrum sp.
0.23%
Figure 3. Decreasing native plant
Lonicera japonica
56.38%
species richness is associated with
Lonicera maackii
0.70%
increasing alien plant coverage.
Rosa multiflora
2.09%
4
2
0
0
10
20
30
+
A. rudis
-
A model of environmental factors significantly
associated with changes in A. rudis abundance. Soil
moisture, tree richness, and abundance were not
associated with A. rudis abundance. Arrow width
represents amount of variance explained by the
association. Arrow direction represents hypothesized
direction of relationship.
Invasive plant coverage is associated with decreasing native plant richness.
In our model of coverage, increasing coverage by one species does not
require decreasing coverage of other species. Coverage was determined by
the presence of a species in a grid of nine subplots. Maximum coverage of a
species at a single site is nine; many species may have a score of 9 at any
one site.
Table 1. Invasive alien plant species
found in this study and their contribution
to the total alien plant species coverage.
6
Trees
Plant Richness
Increased
abundance of A.
rudis is associated
with increased
levels of alien plant
coverage and total
plant species
richness.
References
Aphaenogaster rudis
The aim of this research is to quantify changes in
the arthropod community associated with the
invasion of alien plants in the low forest of Dyke
Marsh Wildlife Preserve (DMWP). Associations
among plant coverage and richness, soil moisture
and structure, tree abundance and richness, and
the arthropod community are examined using
several methods. We hypothesize that increased
alien invasive plant coverage decreases native
plant richness and changes the abundance and
diversity of native arthropods.
B.
Aphaenogaster rudis
Arthropods are ideal for studying changes in
eastern deciduous forests.
Arthropods are
abundant, have short generation times, and are
sensitive to local changes. We have caught and
identified (to various taxonomic levels) nearly
300 species in over 90 families of arthropods in
this study. For this poster we concentrate on the
ant community (Insecta: Hymenoptera: Formicidae). Ants are attractive indicators of change as
they are easy to catch and identify, are diverse,
generally do not travel more than a few meters
from their nests, and move nests away from areas
which are no longer suitable.
A.
Total Ant Abundance
Since colonial times, biodiversity has markedly
changed in the Washington, D.C., area as a result
of disturbances from humans. Unlike direct
human threats such as encroachment, fragmentation, overuse, and pollution, introduced
organisms cannot be stopped by simply passing
legislation or by lawsuits against parties guilty of
introducing these organisms. Alien, invasive
organisms will continue to grow, fill our natural
areas, and consume resources needed to support
our native populations.
10
0
0.0
.2
.4
.6
10
0
.8
1.0
0
2
4
6
8
10
12
14
16
Total Plant Richness
Alien Plant Coverage (%)
Figure 5a. Aphaenogaster rudis
abundance and alien plant coverage.
Figure 5b. Aphaenogaster rudis
abundance and total plant richness.
Conclusions
• Preliminary data show that alien plants are
associated with changes in native plant richness and
ant abundance in the low forest of the Dyke Marsh
Wildlife Preserve. Lonicera japonica and Celastrus
orbiculatus appear to present the larger threats to
native plant species and the terrestrial arthropod
community within the forest.
• Modern computer software and GPS provide an
effective method of randomly sampling a large area
without the use of transects, haphazard sampling,
or other potentially flawed forms of site selection.
Appendix
Table 1. Increased alien plant coverage is
Table 3. Increased alien plant coverage
Crisp, P. N., K. J. M. Dickinson, and G. W. Gibbs.
associated with decreased native plant
is associated with increased A. rudis
1998. Does native invertebrate diversity
richness.a,b
abundance.a
reflect native plant diversity? A case study
from New Zealand and implications for
conservation. Biological Conservation 83:
209–220.
Gremmen, N. J. M., S. L. Chown, et al. 1998. Impact
of the introduced grass Agrostis stolonifera
on vegetation and soil fauna communities at Table 2. Increased alien plant coverage is
Statistical Analysis
Marion Island, sub-Antarctic. Biological
associated with increased ant
Each site’s trap catches are summed
a
abundance.
Conservation 85(3): 223–231.
across all trapping dates (n=40).
Hierarchical Linear Regression was
Panzer, R. and M. W. Schwartz. 1998. Effectiveness
used for analysis where appropriate.
of a vegetation-based approach to insect
Simple regression was used for the
conservation. Conservation Biology 12: 693–
analysis of invasive plant coverage
702.
and native plant richness.
Perfecto, I. and R. Snelling 1995. Biodiversity and the
transformation of a tropical agroecosystem:
Ants in coffee plantations. Ecological
Acknowledgments
Applications 5: 1084–1097.
We greatly appreciate help from Maya Patel, Philip Sze, Martha Weiss,
Coefficientsa
Model Summary
Model
1
R
a
.406
R Sq uare
.165
Adjusted
R Sq uare
.143
Std. Error of
the Estimate
2.164
a. Predictors: (Constant), Invasive
ANOVAb Plant Coverage
Model
1
Reg ression
Residual
Total
Sum of
Squares
35.142
177.958
213.100
df
1
38
39
Mean Square
35.142
4.683
F
7.504
Sig .
.009a
Model
1
2
(Constant)
Total Plant Richness
(Constant)
Total Plant Richness
Alien Plant Coverag e (%)
Unstandardized
Coefficients
B
Std. Error
22.209
4.235
-1.694
.595
14.161
5.859
-1.384
.598
12.162
6.335
Standardized
Coefficients
Beta
-.419
-.342
.284
t
5.244
-2.846
2.417
-2.316
1.920
Sig .
.000
.007
.021
.026
.063
a. Dependent Variable: Aphaenogaster rudis
a. Predictors: (Constant), Invasive Plant Coverag e
b. Dependent Variable: Native Plant Richness
Coefficientsa
Model
1
2
3
(Constant)
Soil Moisture
(Constant)
Soil Moisture
Tree Richness
(Constant)
Soil Moisture
Tree Richness
Percent Alien
Plant Coverag e
Unstandardized
Coefficients
B
Std. Error
54.624
9.919
-126.351
40.526
44.296
9.310
-110.769
36.217
4.175
1.238
32.316
10.797
-90.450
36.319
4.446
1.199
13.552
6.816
Standardized
Coefficients
Beta
-.323
.464
t
5.507
-3.118
4.758
-3.059
3.372
2.993
-2.490
3.707
Sig .
.000
.003
.000
.004
.002
.005
.018
.001
.257
1.988
.054
-.451
-.396
.436
a. Dependent Variable: Total Ant Abundance
John Sauer, Barry Wood (NPS), Smithsonian and USDA specialists, and
support from Friends of Dyke Marsh, Georgetown University, the National
Park Service, and the Washington Biologists’ Field Club.