2011 NORTHEASTERN NATURALIST 18(1):107–114
Does the Seed Bank Reflect Plant Distributions in a
Coastal Dune?
David S. Messina1 and Tara K. Rajaniemi1,*
Abstract - The spatial patterns in the seed bank were examined within a coastal sanddune
system. Soil samples were collected from vegetated and bare plots in three zones,
each dominated by different plant species. The total number of seeds and species diversity
in the seed bank were low. There were no significant patterns in total seed number.
However, there were significant differences in species composition. Seeds of Artemisia
campestris (Tall Wormwood) were more abundant in vegetated plots and were found in
all zones, although adults are more spatially restricted. Fruits of Cyperus grayii (Gray’s
Sedge) were most common where adults are found. Dominant species from the site were
absent from the seed bank. Overall, composition of the seed bank does not reflect composition
of the adult community.
Introduction
A seed bank consists of all the potentially viable seeds contained in the soil
or on the soil surface (Csontos 2007). These seeds provide a source for recruitment
for future generations of plants (Harper 1977). Seed banks may make an
important contribution to plant community structure, particularly in systems that
experience severe, unpredictable disturbances, such as cropland, heathland, or
wetlands that experience extreme flooding (Thompson 1992). While coastal sand
dune communities experience disturbances such as wind, salt spray, and major
storms, very little has been reported on their seed banks.
Baptista and Shumway (1998) studied the seed banks of coastal dunes of Cape
Cod, making comparisons among four sites. They found low abundance and low
species diversity of seedlings emerging from soil samples from all sites. Our
goals in this study were to examine the abundance and species composition in the
seed bank of a coastal dune community at a single site, and to look for within-site
patterns. Baptista and Shumway’s (1998) four sites were all located within Cape
Cod National Seashore (the Lower Cape), whereas our study site is on the Upper
Cape, closer to the mainland.
Patterns in the seed bank may result from patterns in adult vegetation, especially
if most seeds disperse only short distances, as appears to be typical for
coastal dune species (Ehrenfeld 1990). Two reviews of studies comparing the
seed bank to adult vegetation conclude that similarity between adults and seeds is
greatest in frequently disturbed habitats (Hopfensperger 2007, Warr et al. 1993).
The studies reviewed, however, consider the adult vegetation and the seed bank
of an entire site. We examined this question at a smaller scale: given that there
are patterns in distribution of adults within a coastal dune site, does the seed
1Biology Department, University of Massachusetts Dartmouth, North Dartmouth, MA
02747. *Corresponding author - trajaniemi@umassd.edu.
108 Northeastern Naturalist Vol. 18, No. 1
bank exhibit similar patterns? Previous studies of coastal sand dune seed banks
(reviewed by Maun 2009) have not tested for such within-site patterns.
Our study site consists of a barrier dune system with a single dune crest, 1–2 m
in height. Adult vegetation at the site shows strong patterns of zonation with
distance from the shore. Vegetation is patchy, and bare ground is common in all
zones. The purpose of this study was to determine whether the spatial distribution
of species in the seed bank follows the distribution of adult plants. We assessed
the seed bank in the three vegetation zones, and in vegetated and bare plots. We
predicted that more seeds would be found in vegetated than bare plots because
seeds should physically get caught by existing vegetation (e.g., Li 2008). We also
predicted that species composition in the seed bank of each zone would be similar
to the composition of the adult vegetation in that zone.
Field-site Description
Field work was conducted at Waquoit Bay National Estuarine Research
Reserve in Falmouth, Massachusetts (N41°33'5", W70°30'22"). The mean air
temperature and precipitation are 9.8 °C and 1039.5 mm, respectively (averages
from 1949–1970 and 1976–1982, respectively; WorldClimate.com). The study
site is a barrier dune system separating the Atlantic Ocean from Waquoit Bay. The
dune system consists of a single, low dune crest, 2 m or less in height and 5–20 m
from the high tide line. A largely flat area behind the crest extends 100–200 m to
saltmarsh or the bay. The soil throughout the dune system is a coarse sand with
low organic matter content.
The area shows a typical loose zonation of plant species. The windward side
of the dune faces the Atlantic Ocean and is dominated by Ammophila breviligulata
Fern. (American Beachgrass). The leeward side of the dune is dominated by
Rosa rugosa Thunb. (Saltspray Rose). In the flat area behind the dune, Myrica
pensylvanica Mirbel. (Northern Bayberry) is abundant. American Beachgrass is
present at low density in all areas of the dune. Artemisia campestris L. ssp. caudata
(Michx.) Hall & Clem. (Tall Wormwood) is common in open areas of the
dune flats.
Methods
On 22 April 2009, substrate samples were collected in a sequence of paired plots
within the study site’s three zones and along four transects. The sampling date was
expected to be before the beginning of natural germination, as minimum daily temperatures
were near or below freezing prior to this date (United States Historical
Climatology Network, http://cdiac.ornl.gov/epubs/ndp/ushcn/ushcn.html). The
three zones were dominated by American Beachgrass, Saltspray Rose, and Northern
Bayberry. Transects were 120 m apart along the length of the dune. Each of the
four transects contained two sets of plots within each zone, with beachgrass plots
an average of 2 m in front of the dune crest and rose and bayberry plots an average
of 11 and 48 m behind the dune crest, respectively. Each set of plots consisted of a
bare area and a densely vegetated area. In the rose and bayberry zones, vegetated
2011 D.S. Messina and T.K. Rajaniemi 109
plots were located within shrubs, while bare plots were located at least 1 m from
shrubs in areas with little to no vegetation (<5% plant cover in a 1-m2 plot). In
the American Beachgrass zone, completely bare areas were rare; vegetated plots
were located in areas of dense beachgrass growth (>40% plant cover) and “bare”
plots had sparse beachgrass growth (<20% plant cover). A total of 48 plots were
sampled. In each plot, soil was collected from an approximately circular area 15
cm in diameter to a depth of 10 cm. Each core was collected with a garden shovel,
labeled, and bagged.
We used the germination method (Gross 1990) to identify and count seeds in
the seed bank. The 48 samples were each transferred into one half of a divided
plastic tray measuring 25 cm x 25 cm x 6 cm. Each tray was filled with potting
soil approximately 3 cm in depth and topped with the field samples (≈2 cm deep).
The potting soil was used to retain moisture and improve conditions for seedling
germination. Paired bare and vegetated plots were put in halves of the same tray.
The trays were positioned randomly in the greenhouse, with positions rotated
weekly, and watered three times a week.
Every week, germination was recorded. Once a seedling was identified to species,
it was counted and removed from the tray.
The seed count data included many zero values and therefore did not meet the
assumptions of ANOVA. Instead, we used unrestricted permutations of the data
to calculate significance values for the effects of zone, plot type, and their interaction
on total seed number, following the recommendations of Manly (2007).
This analysis was conducted with R version 2.8.1.
We used a χ2 test to assess whether seedling composition was similar among
all zones and plot types. The seedling counts of all dune plant species, other than
Tall Wormwood and Cyperus grayii Torr. (Gray’s Sedge), were grouped together
for the χ2 test, because these species together accounted for less than 4% of the
total seedlings. Thus, the χ2 test compared the distribution of three species (or
species groups) among six zone-plot type combinations. The initial χ2 test was
then broken down into two separate contingency tests for the two most common
species, Tall Wormwood and Gray’s Sedge. These 3 x 2 contingency tests tested
whether zone and plot type had independent effects on seed distribution.
Results
Over the 15-week germination experiment, a total of 388 seedlings germinated
and were identified. Seedlings of Tall Wormwood, Gray’s Sedge, Solidago
sempervirens L. (Seaside Goldenrod), Oenothera parviflora L. (Small-flowered
Evening Primrose), Suaeda maritima (L.) Dum. (Sea Blite), and Lepidium
virginicum L. (Pepperweed) were found. Some additional germinants (67 total)
were of common weed species (Chenopodium album L. [Lambsquarters],
Medicago sativa L. [Alfalfa]) that had never been observed at the field site.
Unfortunately, potting soil controls were not used in this experiment, but these
species had been observed germinating at similar densities in other experiments
using the same potting soil. Therefore, weed seedlings were assumed to have
110 Northeastern Naturalist Vol. 18, No. 1
emerged from the potting soil in the trays and were not included in analyses.
There was little correspondence between the species composition of germinating
seedlings and of adult vegetation in the three zones (Table 1). The three most
abundant species in the adult vegetation for each zone were never observed in the
seed bank.
Tall Wormwood, Gray’s Sedge, and Seaside Goldenrod were the only species
that emerged in samples from the windward (beach) side of the dune
(Fig. 1, Table 1). Wormwood had the highest abundance; its seedling count
in the vegetated plot samples was more than double that in the bare plot
samples. In contrast, Gray’s Sedge had the highest seedling count in the bare
plot samples—double that in the vegetated plot samples. Tall Wormwood,
Table 1. Comparison of adult vegetation and seeds germinated in the three zones. Percent cover for
adult vegetation was measured in 1-m2 plots from 10 transects in July 2009. For dune front, plots
are at the dune crest or 1–11 m in front of the crest (n = 10). For dune back, plots are 1–11 m behind
the dune crest (n = 13). For dune flat, plots are 20-80 m behind the dune crest (n = 19). Only species
with >1% mean cover are shown. Mean and standard deviation of % cover are given, as well
as total number of plots in which the species occurred. For seeds germinated, the total number of
seeds (from n = 8 samples) is given. Bare = bare plot, Veg = vegetated plot.
Adult vegetation
% cover Seeds germinated
Zone / Species mean ± s.d. [# plots] Bare Veg
Dune front
Ammophila breviligulata (American Beachgrass) 22.4 ± 20.5 [9]
Lathyrus japonicus (Beach Pea) 13.9 ± 18.5 [6]
Rosa rugosa (Saltspray Rose) 12.8 ± 30.9 [2]
Toxicodendron radicans (Poison Ivy) 1.9 ± 5.0 [3]
Solidago sempervirens (Seaside Goldenrod) 1.8 ± 2.8 [4] 2 2
Artemisia campestris ssp. caudata (Tall Wormwood) 11 32
Cyperus grayii (Gray’s Sedge) 7 2
Dune back
Saltspray Rose 18.5 ± 29.5 [5]
American Beachgrass 16.2 ± 18.7 [12]
Poison Ivy 9.9 ± 15.9 [6]
Wormwood 6.6 ± 7.0 [8] 30 13
Beach Pea 1.1 ± 2.9 [2]
Lepidium virginicum (Peppergrass) 2
Gray’s Sedge 1
Oenothera parviflora (Small-flowered Evening Primrose) 1
Dune flat
Myrica pensylvanica (Northern Bayberry) 24.0 ± 29.7 [10]
Poison Ivy 11.4 ± 14.8 [11]
Saltspray Rose 7.2 ± 16.4 [4]
Tall Wormwood 5.3 ± 6.4 [11] 17 16
Poa sp. 3.7 ± 6.4 [7]
American Beachgrass 2.5 ± 4.0 [10]
Polygonella articulata (L.) Meisner (Jointweed) 1.4 ± 2.1 [7]
Gray’s Sedge 1.2 ± 2.2 [5] 80 165
Seaside Goldenrod 3 1
Suaeda maritima (Sea Blite) 2 1
2011 D.S. Messina and T.K. Rajaniemi 111
Gray’s Sedge, Small-flowered Evening Primrose, and Peppergrass emerged
in samples from the dune back. Wormwood again had the highest abundance,
but in this zone it was more abundant in bare plots (Fig. 1). Tall Wormwood,
Gray’s Sedge, Seaside Goldenrod, and Sea Blite emerged in samples from the
Figure 1. Mean (± 1
standard deviation)
of seed counts (n =
8) for each plot type
and zone. “Other”
includes Oenothera
parviflora (Smallflowered
Evening
Primrose), Solidago
sempervirens (Seaside
Goldenrod),
Suaeda maritima
(Sea Blite), and Lepidium
virginicum
(Peppergrass).
112 Northeastern Naturalist Vol. 18, No. 1
leeward dune slope. Gray’s Sedge had the highest abundance, and was more
abundant in vegetated plots than bare plots (Fig. 1). Wormwood was the second-
most abundant species with almost equal counts in both plot types.
The total seedling counts (combining all species) were not significantly affected
by zone (P = 0.212 by permutation test), plot type (P = 0.710), or their
interaction (P = 0.899). Results were similar when P-values were derived from
the parametric two-way ANOVA (zone F2,42 = 1.674, P = 0.200; plot type F1,42 =
0.206, p = 0.652; zone*plot type F2,42 = 0.218, P = 0.805). Significant differences
in species composition were found among the six zone-plot type combinations
(χ2 = 221.234, P < 0.001). When examining individual species, we found interacting
effects of zone and plot type on seedling abundance for both Wormwood (χ2 =
16.942, P < 0.001) and Gray’s Sedge (χ2= 8.387, P < 0.001).
Discussion
We found no statistically significant within-site patterns in numbers of seeds
in the germinable seed bank, either among zones or plot types. The dune flat
samples tended to have higher seed numbers, due to a few samples with high densities
of Gray’s Sedge (note the high standard deviation in Fig. 1). Adults of this
species occur in isolated patches (T.K. Rajaniemi, unpubl. data); if most achenes
disperse only short distances, samples with high numbers of sedge fruits may
have been collected near patches of adults. The windward and leeward slopes of
the dune had similar seed densities to each other.
The presence of dense grass or shrubs also did not affect seed density. We had
hypothesized that, in the windy dune environment, vegetation would physically
trap seeds (Li 2008). Several studies in desert systems have found that shrubs are
associated with either higher or lower seed numbers in the seed bank than gap
areas. These studies indicate the importance of local dispersal: when annuals are
more abundant under shrubs, seeds are also more numerous under shrubs (Guo
et al. 1998, Holzapfel et al. 2006), and when annuals are more abundant in gaps,
seeds are more numerous in gaps (Holzapfel et al. 2006, Sternberg et al. 2004). In
our system, the species represented in the seed bank are found primarily in gaps,
but seeds are equally numerous under shrubs, suggesting that the wind may move
seeds over distances of meters to tens of meters.
We did find some patterns in species composition of the seed bank. These
patterns can be attributed to the two most common species in the seed bank, Tall
Wormwood and Gray’s Sedge. Tall Wormwood, a biennial herb, is a pioneer
species that colonizes areas of major wind disturbance (C. Weideman, Waquoit
Bay NERR, Falmouth, MA, pers. comm.). It is found on the leeward slope of the
dune and particularly in the dune flat, but rarely under shrubs and never on the
windward slope of the dune (Table 1; T.K. Rajaniemi, unpubl. data). Its seeds
were found in all vegetation zones, and were associated more with dense vegetation
on the windward slope and with bare areas on the leeward slope. Wormwood
was also the most common species found in a previous study of Cape Cod sand
dune seed banks, but its distribution was clumped both between and within sites
2011 D.S. Messina and T.K. Rajaniemi 113
in that study (Baptista and Shumway 1998). The wide spatial distribution of Tall
Wormwood seeds may reflect a ruderal strategy. We have yet to determine why
adult Tall Wormwood are not present on the dune front, where seeds are found.
As noted above, Gray’s Sedge adults are found only in the dune flat, where its
fruits are also most common. A few fruits of this species experience some wind
dispersal, and are found in the dune front, 50–100 m from the adults. Adults are
often found near Northern Bayberry shrubs, and these appear to capture windblown
seeds, so that seed numbers are higher under shrubs in the dune flat.
Species diversity in the seed bank was low, and the dominant species on
the dune were not represented in the germinable seed bank. Previous studies
have also observed low diversity in coastal dune seed banks (Baptista and
Shumway 1998, Looney and Gibson 1995). Baptista and Shumway (1998)
found mostly subordinate species in the seed bank, and Looney and Gibson
(1995) noted that most woody species were absent from the seed bank. Dominant
woody species including Saltspray Rose, Bayberry, and Toxicodendron
radicans (L.) Kuntze. (Poison Ivy), as well as herbaceous species such as
Lathyrus japonicus Willd. (Beach Pea), produce copious seeds at our site, and
these seeds germinate well in the greenhouse (Poison Ivy germination has not
been tested; T.K. Rajaniemi, unpubl. data). Some of these seeds may be lost
in the field to predators; we have observed weevil damage in about half of the
Beach Pea seeds collected and noted many chewed seeds of Saltspray Rose
on the ground while collecting soil samples, suggesting fruit and/or seed consumption
by small mammals. Short seed life span, attack by pathogens, deep
burial of seeds, or transport out of the dune system might also account for
some seed loss (Baptista and Shumway 1998). In general, seed banks of coastal
sand dunes appear to be largely transient, although the fates of dispersed
seeds have not been well studied (Maun 2009).
In conclusion, the species composition of the germinable seed bank at this
coastal dune site did not reflect the species composition of the adult vegetation,
for the site as a whole or along gradients within the site. The seed bank may
contribute to the ability of Tall Wormwood to colonize blowouts, but it does not
appear to contribute to maintenance of the dominant vegetation.
Acknowledgments
We wish to thank Chris Weideman and the Waquoit Bay National Estuarine Research
Reserve for advice and access to the study site.
Literature Cited
Baptista, T.L., and S.W. Shumway. 1998. A comparison of the seed banks of sand dunes
with different disturbances histories on Cape Cod National Seashore. Rhodora
100:298–313.
Csontos, P. 2007. Seed banks: Ecological definitions and sampling considerations. Community
Ecology 8:75–85.
Ehrenfeld, J.G. 1990. Dynamics and processes of barrier island vegetation. Reviews in
Aquatic Sciences 2:437–480.
114 Northeastern Naturalist Vol. 18, No. 1
Gross, K.L. 1990. A comparison of methods for estimating seed numbers in the soil.
Journal of Ecology 78:1079–1093.
Guo, Q., P.W. Rundel, and D.W. Goodall. 1998. Horizontal and vertical distribution of
desert seed banks: Patterns, causes, and implications. Journal of Arid Environments
38:465–478.
Harper, J.L. 1977. Population Biology of Plants. Academic Press, London, UK. 892 pp.
Holzapfel, C., K. Tielbörger, H.A. Parag, J. Kigel, and M. Sternberg. 2006. Annual plantshrub
interactions along an aridity gradient. Basic and Applied Ecology 7:268–279.
Hopfensperger, K.N. 2007. A review of similarity between seed bank and standing vegetation
across ecosystems. Oikos 116:1438–1448.
Li, F.-R. 2008. Presence of shrubs influences the spatial pattern of soil seed banks in
desert herbaceous vegetation. Journal of Vegetation Science 19:537–548.
Looney, P.B., and D.J. Gibson. 1995. The relationship between the soil seed bank and
above-ground vegetation of a coastal barrier island. Journal of Vegetation Science
6:825–836.
Manly, B.F.J. 2007. Randomization, Bootstrap, and Monte Carlo Methods in Biology, 3rd
Edition. Chapman and Hall, London, UK. 480 pp.
Maun, M.A. 2009. The Biology of Coastal Sand Dunes. Oxford University Press, New
York, NY.
Sternberg, M., S.L. Yu, and P. Bar. 2004. Soil seed banks, habitat heterogeneity, and
regeneration strategies in a Mediterranean coastal sand dune. Israel Journal of Plant
Sciences 52:213–221.
Thompson, K. 1992. The functional ecology of seed banks. Pp. 231–258, In M. Fenner
(Ed.). Seeds: The Ecology of Regeneration in Plant Communities. CAB International,
Oxon, UK.
Warr, S.J., K. Thompson, and M. Kent. 1993. Seed banks as a neglected area of biogeographic
research: A review of literature and sampling techniques. Progress in Physical
Geography 17:329–347.