Download alteration of ecosystem nitrogen dynamics by exotic plants: a case

INVASIVE SPECIES AND THE SOIL
October 2001
1323
Ecological Applications, 11(5), 2001, pp. 1323–1335
� 2001 by the Ecological Society of America
ALTERATION OF ECOSYSTEM NITROGEN DYNAMICS BY EXOTIC
PLANTS: A CASE STUDY OF C4 GRASSES IN HAWAII
MICHELLE C. MACK1,3 CARLA M. D’ANTONIO,1
AND
RUTH E. LEY2
1Department
of Integrative Biology, University of California, Berkeley, California 94720 USA
Department of Environmental, Organismic, and Population Biology, University of Colorado,
Boulder, Colorado 80309 USA
2
Abstract. Biological invaders can alter ecosystem processes via multiple pathways, yet
few studies have compared the relative importance of these pathways. We assessed the
impacts of exotic, invasive grasses on ecosystem nitrogen (N) cycling in the seasonal
submontane woodlands of Hawaii Volcanoes National Park, where native grasses have been
historically rare. Exotic grasses have become abundant over the past 30 yr and have altered
two controls over N cycling: plant species composition and fire regime. Here we synthesize
the results of a long-term investigation of species impacts in this system. To determine
effects of grasses and fire on internal N cycling, we compared litterfall, decomposition, N
mineralization from soil organic matter (SOM), and plant N uptake and production in
invaded unburned forest, grass-removal plots within the forest, and woodland converted to
grassland by fire. We measured ecosystem N loss via fire by comparing N pools among
unburned, naturally burned, and experimentally burned sites. We also assessed the effects
of fire on annual N fixation in the unburned forest vs. the grassland.
Exotic grasses had relatively small effects on N cycling in the unburned woodland
despite being abundant in the understory for 30 yr. Grasses contributed �30% of fine
litterfall and primary-production mass and N in the unburned woodland. However, these
contributions did not result in significantly increased totals because litterfall and production
of Metrosideros polymorpha, the dominant native tree, was reduced in the invaded woodland
relative to grass-removal plots, presumably due to competition with grasses. Although areaweighted decomposition was lower in the grass-removal treatment than in the control, net
N mineralization from litter and SOM were similar between these treatments. Annual plant
N uptake was similar to annual net N mineralization from SOM in both treatments.
By contrast, the burned grassland exhibited much lower rates of litterfall and production
mass and N, but higher rates of net N mineralization from SOM than the woodland. As a
result, total annual plant N uptake was only 17% of annual net mineralization. This change
was primarily due to the loss of native species. Aboveground N pools were significantly
reduced with fire. Native species were largely eliminated by fire. However, across all burned
and unburned sites there was no change in total ecosystem N because the N contained in
biomass was relatively small compared to N in litter and soil. Soil contained �95% of
ecosystem N in all sites. Only in the high-intensity experimental burn was there significant
loss of N from the soil pool. Fire reduced N inputs through asymbiotic N fixation mainly
due to the loss of M. polymorpha, whose litter is an important site of asymbiotic N fixation,
and alteration of the soil O-layer. This reduction in N inputs makes it unlikely that fixation
activity will replace N lost via combustion before the next fire. Fire and the ensuing loss
of native species led to decreased N inputs, increased rates of N mineralization from litter
and SOM, and dramatically reduced plant N uptake, potentially leading to a more leaky N
cycle. It appears that the indirect effects of grasses on N cycling via the elimination of
native species by fire is the most important pathway though which exotic grasses alter
ecosystem N dynamics in this system.
Key words: biological invasions; decomposition; ecosystem nitrogen dynamics; exotic grasses;
fire; Hawaii Volcanoes National Park; litterfall; nitrogen fixation; nitrogen loss; plant nitrogen uptake;
primary production; soil nitrogen mineralization.
Manuscript received 17 September 1999; accepted 26 August 2000; final version received 5 October 2000. For reprints of this
Invited Feature, see footnote 1, p. 1259.
3 Present address: Institute of Arctic Biology, University of Alaska, Fairbanks, Alaska 99775-7000 USA.
E-mail: [email protected]
1323
INVITED FEATURE
1324
INTRODUCTION
Although ecologists have acknowledged for decades
that plant species composition can influence ecosystem
properties, it has only been in the past two decades
that ecosystem scientists have begun to explore the
mechanisms through which these effects are exerted.
Species composition has been demonstrated to influence such ecosystem characteristics as nitrogen accumulation and cycling (Boring and Swank 1984, Vitousek et al. 1987, Wedin and Tilman 1990, Hooper and
Vitousek 1998), carbon storage (Fisher et al. 1994,
Wedin and Tilman 1990, Schlesinger and Pilmanis
1998, Christian and Wilson 1999), water availability
(Holmes and Rice 1996) and run-off (Le Maitre et al.
1996), and disturbance regime (D’Antonio and Vitousek 1992, Mack and D’Antonio 1998). These effects
can incur direct economic costs such as reduced water
availability for human consumption (VanWilgen et al.
1996), or they may influence the trajectory of community and ecosystem development. Without an understanding of key mechanisms, ecologists cannot predict how rapidly species introductions may cause ecosystem-level change, and the potential for management
to reverse these changes.
Nitrogen cycling is an ecosystem function that may
be particularly sensitive to changes in species composition. Species may affect ecosystem nitrogen (hereafter, N) cycling and storage by influencing rates of N
input (e.g., N fixation; Vitousek et al. 1987, Musil and
Midgley 1990, Witkowski 1991, Stock et al. 1995, Maron and Jefferies 1999), the amount and timing of N
uptake (Joffre 1990, Tilman et al. 1996), chemistry and
quantity of litter material entering the soil organic matter (SOM) pool (Vitousek 1982, Wedin and Tilman
1990), and/or microclimate conditions in which decomposer activity occurs (Kochy and Wilson 1997,
Breshears et al. 1998). Species effects on N cycling
have been relatively well documented for N-fixing species because individual species have strong, clear effects on both N inputs and rates of internal cycling
(Vitousek et al. 1987, Vitousek and Walker 1989, Stock
et al. 1995, Maron and Jefferies 1999). For other woody
species, leaf-litter chemistry is often strongly correlated with decomposition rates that in turn correlate
with soil net N mineralization (Scott and Binkley
1997). Relatively few studies have established the influence of herbaceous species on ecosystem N cycling
and storage when these species are a part of a system
dominated by woody species (but see Berendse et al.
[1987] [1989], Asner and Beatty [1996]).
Biological invasion, the successful human-caused
establishment and spread of non-native species into
native communities, has resulted in numerous opportunities to examine species effects on ecosystem N cycling. Currently we have no overarching framework for
assessing which species or species traits are likely to
Ecological Applications
Vol. 11, No. 5
have strong ecosystem impacts, or the circumstances
under which individual species will cause rapid or detectable changes. Chapin et al. (1996) suggested that
species that will affect ecosystem processes are those
that have traits that are qualitatively different from resident species. D’Antonio et al. (1999) suggested that
these species may have a trait or suite of traits relative
to a particular ecosystem function that have no functional analog in the traits of native species in the invaded system. By contrast, species that differ from natives mainly in traits that vary continuously among
species in the community, such as size, relative growth
rate, or tissue quality, would be expected to overlap
with native species in their effects on resources, and
thus interact with residents in a competitive manner
that would likely result in similar effects on ecosystem
processes. We suggest that these species can still cause
ecosystem change, but that it will occur much more
slowly. To be useful, this simplistic framework must
be used in reference to clearly defined ecosystem processes or functions.
In Hawaii Volcanoes National Park, introduced C4
grasses have invaded the understory of seasonally dry
submontane woodlands where they compete with native species for limiting resources (D’Antonio et al.
1998). Native graminoids are rare in these woodlands
(Smith and Tunison 1992), thus invasion has introduced
an herbaceous perennial growth form into an ecosystem
previously dominated by woody, evergreen species.
Exotic grasses have functionally different fuel characteristics from native species because they form a continuous understory canopy where standing dead biomass with a high surface-area-to-volume ratio is always
high. By contrast, the understory canopy of native species is patchy, their litter drops to the ground, and very
little standing dead native material is present in the
understory. As a result of this new suite of traits related
to fuel loading, fire frequency and size have increased
dramatically in invaded sites (Smith and Tunison 1992,
D’Antonio et al. 2000). After fire, introduced grasses
dominate these sites for decades (Hughes et al. 1991).
Primary production in nearby seasonal submontane
forests (Vitousek et al. 1995) and the grass-dominated
ecosystems that follow fire (C. M. D’Antonio, unpublished data) has been experimentally determined to be
N limited. Thus, changes in plant species composition,
N inputs, or N loss may have important consequences
for N cycling and primary production. Grasses may
alter ecosystem N cycling directly through their use of
N, including litter production, decomposition, SOM
formation, and N uptake and sequestration. Traits that
control plant N use vary continuously among native
and exotic species; every species contributes some litter with concentrations of carbon (C) and nutrients that
overlap among species (Chapin et al. 1996). If grasses
alter litterfall mass, litter decomposability, or amount
of N taken up by biomass, they may shift the ratio of
INVASIVE SPECIES AND THE SOIL
October 2001
C to N cycling in the system, resulting in alteration of
N availability to primary production. Furthermore,
grasses may alter ecosystem N cycling indirectly via
their effects on fire, including direct combustion of
biomass and loss of N to the atmosphere (e.g., Raison
et al. 1985, Kauffman et al. 1993), and removal of
native species (Hughes and Vitousek 1993, D’Antonio
et al. 2000). The invasion of grasses into these woodlands prior to fire and the rise to complete dominance
by these grasses after fire offer a unique opportunity
to investigate the impact of an herbaceous species both
on ecosystem processes in the prefire woody community, and on the ecosystem after the dramatic compositional changes induced by fire.
Over the past eight years we have investigated the
impact of introduced grasses and fire on growth and
production of woody species and N cycling in submontane woodlands. We demonstrated strong competitive interactions among grasses and native woody species for limiting resources including N in the unburned
woodland (D’Antonio et al. 1998). Because interactions were competitive, we predicted that grass effects
on N dynamics prior to fire would be small because
the traits that control N cycling overlap between grasses and the native species. By contrast, we predicted
that impacts of grasses on N dynamics via fire would
be relatively large because in this case grasses have
introduced a fundamentally new process (fire) to the
ecosystem and triggered a large change in species composition. The data reported in this manuscript are part
of a long-term investigation of species impacts in this
system and represent a synthetic overview of that work.
Study system
Our study sites are within the seasonal submontane
zone of Hawaii Volcanoes National Park (HVNP) on
Kilauea volcano on the leeward side of Hawai’i Island.
They lie within a broad band of open-canopy forest
dominated by the evergreen tree Metrosideros polymorpha (19�6� N, 155�33� W). Since grass invasions
began in the 1960s, fires intruding from campgrounds
and roads have extended into the forest, creating a mosaic of burned and unburned forest patches. Our sites
are around 870-m elevation on pahoehoe lava flows
between 500 and 1000 yr old (Holcomb 1987). Soils
are predominantly entisols developed from volcanic
ash deposited over the past 500 yr. Soil depth ranges
from �10 cm to �1 m; including rocky outcroppings
it averages 30–35 cm (C. M. D’Antonio, unpublished
data). The area receives a mean of 1500 mm of rain
per year with a pronounced drought from April through
September.
The understory of unburned woodland is dominated
by native, evergreen shrubs including Styphelia tameiameiae, Dodonaea viscosa, Osteomeles anthyllidifolia, and Wikstroemia phylliraeifolia. Three species
of native sedges are also present at low abundance.
1325
Schizachyrium condensatum, a grass from tropical
America, makes up 30% of the aboveground understory
biomass and �60% cover (Hughes et al. 1991). Two
other introduced grasses, Andropogon virginicus and
Melinis minutiflora, make up �8% cover, and native
grasses are rare. To simulate conditions that might have
existed in the woodland prior to grass invasion, we
experimentally removed grasses from four replicate 20
� 20 m plots in 1991 and at the same time designated
control plots (D’Antonio et al. 1998). This treatment
was necessary because there are no comparable areas
in Hawai’i that lack introduced grasses.
Burned sites are dominated primarily by M. minutiflora and S. condensatum and, with the exception of
D. viscosa, native trees and shrubs are absent. Melinis
minutiflora cover increases from 7.2% in unburned
woodland to 79.3% in twice-burned sites, and livegrass biomass increases from �10 g/m2 to �700 g/m2
(C. M. D’Antonio, unpublished data).
METHODS
Impacts of exotic grasses and fire
on internal N cycling
We examined both the direct effects of grasses on N
cycling in the unburned forest, and the indirect effects
of grasses on N cycling via compositional change induced by fire. We estimated annual fine litterfall, litter
decomposition, net N mineralization from litter and soil
organic matter, plant N uptake, and net primary productivity on an area-weighted basis in three sites/treatments with the following vegetation histories: (1) unburned woodland with grasses removed (hereafter
‘‘Woodland’’), (2) unburned woodland with grasses
present (hereafter ‘‘Woodland � Grass’’), and (3)
woodland that had burned in both 1970 and 1987 (hereafter ‘‘Grassland’’). Measurements were made in four
20 � 20 m permanent plots in each site. Woodland and
Woodland � Grass plots were spatially interspersed
and the removal treatment was randomly applied
(D’Antonio et al. 1998), while Grassland plots were
located in burned areas adjacent to unburned forest in
a stratified random manner (Mack 1998). All plots were
selected to have similar aspect and soil depth, and are
located on the same lava flow (Wolf and Morris 1996).
Pre-fire aerial photos show that all plots were part of
a once-continuous Woodland belt (Hawaii Volcanoes
National Park [HVNP] Resource Management Division
Archives).
Litterfall and decomposition.—In each plot, aboveground fine litterfall and N content was measured with
eight 0.25-m2 litter traps (for natives; Crews et al. 1995)
and destructive harvests (for grasses and sedges; Mack
1998) for 12 mo during 1994–1995. We used an in situ
litterbag decomposition experiment (Swift et al. 1979)
to estimate mass and net N loss rate constants for each
common species in each site. Leaves, stems, and twigs
were collected from senesced but still attached litter in
1326
INVITED FEATURE
each site in December 1994 and dried at 30�C for 48
h, after which 1-g increments were sewn into 10 � 10
cm bags of 1-mm mesh fiberglass. Eight replicate litterbags of each species were placed into their plot of
origin in January 1995 and two bags of each litter type
were collected from each plot at 6, 12, 18 mo, and two
years after deployment. At each collection, bags were
cleaned, dried, ground, and analyzed for percentage N
at the University of California (Berkeley California on
a Carlo Erba NA 1500 CHN autoanalyzer (Fisons Instruments, Beverly, Massachusetts, USA). Indices of
mass and N loss from annual litterfall were calculated
for each species as the rate of mass or N loss during
year 1 multiplied by annual litterfall mass or N plus
the rate of mass or N loss during year 2 multiplied by
the residual litter mass or N remaining after one year
of decomposition, then summed across species (M. C.
Mack and D’Antonio, unpublished manuscript).
To compare the inherent decomposability (litter quality) of native species to that of exotic species, we decomposed leaf litter from the dominant native species,
leaf and stem litter from exotic grasses, and twigs from
the native shrub that is present in all sites (Dodonaea
viscosa), in a common garden. Leaves, stems, and twigs
were collected from the Woodland � Grass plots in December 1994, treated and collected as described above,
but placed into a common area in the unburned woodland
for incubation. Here we report percentage of initial mass
remaining at the two-year time point. In conjunction
with this common-garden experiment, we also examined
treatment effects on litter quality within the dominant
tree species. Here, we decomposed Metrosideros polymorpha litter that had been collected in either the Woodland or the Woodland � Grass treatment in the common
garden (M. C. Mack and D’Antonio, unpublished manuscript).
We analyzed vegetation site effects on total fine litterfall, mass, and N loss indices, and common-garden
mass and N loss using one-way ANOVA and t tests.
If necessary, data were ln-transformed to meet the assumptions of homogeneity of variance. We used Tukey
tests for post hoc multiple comparisons in all tests. This
protocol was used for all analyses described hereafter.
All statistical analyses were performed using Systat 7.0
(SYSTAT 1997).
Soil net N mineralization.—We used intact, resincapped soil-core incubations (Hart and Firestone 1989)
to estimate net N mineralization from soil organic matter, our proxy for plant available N. Four cores were
deployed in each replicate plot of each site every 2 mo
for 12 mo during 1994–1995. Pre- and post-incubation
soils were analyzed for ammonium and nitrate concentrations, and net mineralization was calculated as
the difference between post-incubation and pre-incubation ammonium plus nitrate. To calculate annual net
N mineralization, we summed N mineralized in cores
Ecological Applications
Vol. 11, No. 5
over the 12-mo period and tested for differences among
means with ANOVA.
Primary production and N uptake.—We assessed
aboveground plant production and N uptake in each
site using a combination of regression and inventory
methods (shrubs and trees; D’Antonio et al. 1998,
Mack 1998) and destructive harvests (grasses). Allometric equations for shrubs and trees were developed
specifically for this site and are reported in D’Antonio
et al. (1998). Annual litter production (described
above) was summed with bole or with woody or herbaceous perennial stem growth increment to estimate
aboveground net primary production (NPP). To estimate total N uptake for each species, we multiplied
production values by tissue nitrogen concentration
(hereafter ‘‘[N]’’) content averaged over four annual
sampling dates for each tissue type (Mack 1998).
We used a sequential coring method to measure belowground production in the Woodland � Grass and
Grassland sites (Caldwell and Virginia 1989, Vogt and
Persson 1991). Cores 20 cm deep were taken at 3-mo
intervals over the course of one year at 25 randomly
selected locations within the Woodland � Grass and
Grassland treatments. Roots were extracted from the
cores at University of California at Berkeley using a
Bel-Arte soil elutriator (Bel-Arte, Pequannock, New
Jersey, USA). Non-root organic matter was separated
from the cleaned root mass from each core by hand
and roots were then separated into size fractions (� or
�2 mm) and live vs. dead and shrub/tree vs. grass roots.
For the Woodland � Grass site, statistically significant
differences between time points were used to create an
annual production estimate (Caldwell and Virginia
1989). This value was compared to that obtained using
the difference between the annual maximum and minimum values (hereafter ‘‘max-min’’). We report both
values here. To estimate root production in the Woodland, we multiplied aboveground shrub and tree production in that treatment by the root production:shoot
production ratio for shrubs and trees in Woodland �
Grass. This calculation is reported for both significant
difference and max-min estimates to represent the
range of potential root-production values. Grassland
root production was calculated by treating differences
between adjacent time points as significant if any of
the sorted components of total root biomass were significantly different. The annual production value obtained in this way was identical to that obtained using
a max-min estimate. To determine N uptake associated
with belowground production in Woodland and Woodland � Grass, we multiplied significant-difference rootproduction estimates by root [N] for live grass or native
roots. For the Grassland, we multiplied max-min root
production estimates by root [N] for live grass or native
roots. Root [N] were averaged over four sampling dates
between 1995 and 1996 (Mack 1998).
October 2001
INVASIVE SPECIES AND THE SOIL
Impact of grass-fueled fire on ecosystem N pools
To evaluate the effect of fire on N loss we assessed
biomass and soil N pools in four sites reflecting different burn histories: (1) unburned woodland surrounding the Woodland � Grass plots described above (see
Impacts of exotic grasses and fire. . . ), (2) woodland
that had burned once in 1970 during one of the first
large grass-fueled fires in HVNP, (3) woodland that had
burned once in 1987, and (4) woodland that had burned
in both 1970 and 1987 (surrounding the Grassland plots
described above). Aerial photos show that all sites were
part of a once-continuous forest belt that has been
carved up by fire beginning in 1970. Understory biomass was sampled during July and August 1992 by
harvesting all aboveground plant parts in 2 � 2 m plots
from each of 10 locations within each site. Plant material was transported to the laboratory and then separated by species and tissue type. Sorted material was
then weighed moist, and subsamples were dried at 70 �C
for estimation of dry mass and percentage N. Trees
were inventoried in four, 10 � 100 m transects through
each site and basal diameter and height measurements
were used in conjunction with allometric equations to
estimate tree biomass. Allometric equations were developed specifically for these sites and are reported in
D’Antonio et al. (1998). Root N was assessed by washing and hand separating roots from two 20-cm-deep
cores in each plot. A second set of volumetric 20-cmdeep cores was taken for estimation of soil N pools
and bulk density. Soils were sieved to 2 mm, roots were
removed with forceps, soils were dried at 70�C, ground,
and percentage N was analyzed on the autoanalyzer at
the University of California, Berkeley. Because soil
bulk density did not differ among sites (data not presented), N pools were calculated based on percentage
N per gram soil times mass of soil in a 1 � 1 � 0.2
m cube using the average field bulk density across all
sites (0.63 � 0.11 g/cm2 [mean � 1 SE]).
The above chronosequence approach to N loss confounds direct N loss due to combustion with post-fire
N loss due to erosion, leaching, or gaseous losses (Dryness et al. 1983, Kauffman et al. 1998). We isolated
direct N loss from fire by evaluating N loss immediately
following a low- and a high-intensity experimental
burn. Both fires were set in areas that had burned once
in 1970 and were now grass dominated. Their composition resembled that of our Grassland sites described
above. The low-intensity fire was an �6-ha backing
fire and was ignited in January 1993. It was part of the
forest belt described above. Biomass and N pools were
evaluated as above although sorted material was
weighed in the field after which it was returned to the
sites. A subsample was returned to the laboratory for
dry-mass correction analysis. The high-intensity fire
was a head fire of similar size and was ignited in June
1995 under extreme fire conditions. It was located at
a similar elevation but in a slightly wetter site sur-
1327
rounded by recent (1970) lava flows from Mauna Ulu.
Both sites were dominated by Schizachyrium condensatum and Melinis minutiflora with scattered individuals of Dodonaea viscosa. Pre- and immediate postfire
measurements are reported here.
Impact of fire on potential for biological N fixation
We examined the effects of fire on the potential for
fixation of atmospheric N by first surveying all potential substrates for nitrogenase activity in the unburned
woodland surrounding the Woodland plots and in the
twice-burned woodland surrounding the Grassland
plots. The results of this survey and rates of acetylene
reduction per substrate are presented in Ley and
D’Antonio (1998). Here, we use rates and conversion
factors provided in Ley and D’Antonio multiplied by
measured volumes of the various substrates and an estimated 300 d per year over which fixation can occur
to obtain an estimate of total N inputs per year per site.
Our values differ from those reported in Ley and
D’Antonio (1998) for several reasons: (1) they have a
multiplication error in their calculation of total fixation
in the unburned woodland, (2) we used different estimates of substrate biomass based on values reported
in Mack (1998), and (3) we used C2H2 production estimates based on an ambient temperature of 18�C rather
than 28�C. We also averaged the fall and spring values
reported in Ley and D’Antonio (1998). In addition to
this survey, we assayed soil and root material for nitrogenase activity after fire in the first controlled burn.
RESULTS
AND
DISCUSSION
Effects of grasses on internal N cycling before fire
The invasion of grasses into the unburned woodland
resulted in a dramatic change in the composition of fine
litterfall, but no change in the total amount of litterfall
on an annual basis (Fig. 1). Grasses contributed 34%
of fine litterfall in the Woodland � Grass site, but this
was compensated for by a significant decrease in the
litter contribution of the native tree Metrosideros polymorpha in Woodland � Grass relative to the Woodland (t1,3 � �2.43, P � 0.05; Fig. 1A). Nitrogen delivered in litterfall similarly showed compositional
change, but no change in total amount (Table 1). This
compensatory response indicates competitive suppression of M. polymorpha by grasses, as was demonstrated
in a growth study at the same sites, where we concluded
that increased growth in Woodland natives relative to
Woodland � Grass natives was at least partially due
to a release from grass competition for N (D’Antonio
et al. 1998). We ruled out competition for water because
soil moisture was consistently lower and plant-level
indicators of water stress were consistently higher in
the Woodland treatment than in the Woodland � Grass.
Decomposition on an areal basis was faster in Woodland � Grass than in the Woodland (Fig. 1B). This was
due to two factors: (1) higher decomposability of Schi-
Ecological Applications
Vol. 11, No. 5
INVITED FEATURE
1328
FIG. 1. (A) Mean annual aboveground fine litterfall, (B) litterfall mass loss, and (C) aboveground net primary productivity
(ANPP) in W (Woodland grass-removal treatment), W�G (Woodland � Grass control), and G (burned grassland) sites in
the seasonal submontane zone of Hawaii Volcanoes National Park. Metrosideros polymorpha is the dominant native tree.
Bars with the same lowercase letter above are statistically indistinguishable at P � 0.05 using a Tukey multiple-comparison
test.
zachyrium condensatum stem litter (the dominant exotic grass in the unburned woodland) compared to most
native litters, and (2) increased decomposition of native
litters in the presence of grasses. In the common-garden
experiment, grass stems lost mass and N more quickly
than all other litters (for mass, F6,20 � 28.25, P �
�0.001; for N, F6,20 � 8.69, P � �0.001, Fig. 2) except
Dodonaea viscosa, a native species that contributed
�1% of native litter. Grass stems made up 61 � 22%
(mean � 1 SE) of grass litter production and decomposed faster than the M. polymorpha leaf litter they
replaced. Increased decomposability of native litter was
primarily accounted for by increased initial litter qual-
ity of M. polymorpha leaf litter in Woodland � Grass
relative to Woodland (t6 � �3.50, P � 0.01) since
values for mass loss of M. polymorpha litter in the
common garden were similar to values from the in situ
experiment (Fig. 3).
Net N immobilization on decomposing litter was
somewhat faster in the Woodland � Grass than in the
Woodland, but this difference was only marginally significant (P � 0.11; Table 1). Immobilization dominated
litter net N flux over the two-year period; we expect
that lower rates of immobilization will correlate positively with higher rates of N loss in the following years
(Paul and Clark 1989). Net N mineralization from SOM
TABLE 1. Nitrogen flux data and ANOVA results for annual flux of nitrogen within ecosystem
compartments in three vegetation treatments within the seasonal submontane zone of Hawaii
Volcanoes National Park.
Treatments ‡
N flux (gN·m�2 ·yr�1)†
Litterfall
Litterfall N mineralization
SOM N mineralization§
Plant uptake
Aboveground
Belowground�
Total
W�G
W
1.05
(0.15)
�0.28
(0.10)
3.57a
(1.31)
1.50a
(0.18)
4.0
5.5
a
1.06a
(0.11)
�0.09
(0.07)
4.08a
(1.07)
1.21a
(0.19)
3.2
4.4
G
F
P
0.25b
(0.06)
�0.10
(0.04)
11.10b
(3.80)
0.31b
(0.01)
1.6
1.9
19.07
�0.001
2.79
0.11
4.28
0.05
16.30
0.004
Notes: For flux calculations see Methods: Impacts of exotic grasses and fire on internal N
cycling. Lowercase letters are statistically different at P � 0.05 using a Tukey comparison test.
† There were n � 4 plots except for belowground production, where n � 25 random locations.
‡ W � Woodland (unburned woodland with grasses removed), W�G � Woodland � Grass
(unburned woodland invaded by grasses), and G � Grassland (woodland burned twice and
converted to grassland).
§ SOM � soil organic matter.
� Statistics were not calculated due to estimation methods. For biomass ranges, see Results:
Effects of grasses on internal N cycling before fire.
October 2001
INVASIVE SPECIES AND THE SOIL
1329
row et al. 1990, Nadelhoffer and Raich 1992). Furthermore, N uptake was similar or greater than N supply in both of these sites (Table 1). The cycling of N
between vegetation and soil appears to be quite similar
in these two ecosystems despite the abundance of exotic grasses, the higher decomposability of grass stem
litter, and the increased decomposability of M. polymorpha litter. These changes in litter quality may take
longer to translate to altered rates of N mineralization
from SOM. One reason may be that 95% of ecosystem
N resides in the soil in this system and soils are high
in percentage C (12 � 2% in top 15 cm [mean � 1 SE]
Mack 1998). In general, young ash-based soils have a
large capacity to ‘‘fix’’ organic matter and sequester
soil C (Torn et al. 1997), and thus may respond slowly
to perturbations in C and N inputs. Similarly, Berendse
et al. (1989) concluded that field-based estimates of N
mineralization under invading grasses differed from
modeled estimates because the larger pool of older organic matter dominated N transformations. In these
cases, where invaders differ from natives in traits that
are distributed continuously among species (e.g., litter
quality and production), invaders have little impact on
N cycling at least on the time scale of �20 yr.
Effects of grasses and fire on internal N cycling
Both annual fine litterfall and litterfall N content
were significantly reduced in the Grassland relative to
the unburned woodland (Fig. 1A and Table 1). These
reductions were due to the loss of natives, since grass
contributions were the same whether grasses were
alone (Grassland) or with natives (Woodland � Grass;
Fig. 1A) even though there was a change in grass spe-
FIG. 2. Percentage of initial mass remaining after two
years of decomposition for native and exotic leaf (L) and
stem (S) litters decomposed in litterbags in a common garden
in the seasonal submontane zone of Hawaii Volcanoes National Park. Data are means � 1 SE; n � 4 replicates. Bars
with the same lowercase letter are statistically indistinguishable at P � 0.05 using a Tukey multiple-comparison test.
(soil organic matter) was not different between the two
treatments (Table 1).
Despite a compositional change within aboveground
production that was similar to the change in litter production (Fig. 1C), the Woodland and Woodland �
Grass sites were very similar in total plant production
and N uptake (Table 1). Belowground production
ranged from 362 to 504 g·m�2·yr�1 in Woodland �
Grass (significant differences and max-min method, respectively), from 452 to 643 g·m�2·yr�1 in the Woodland treatment. Root-production values were within
bounds found in other forested ecosystems (Kumme-
FIG. 3. Mean (� 1 SE) percentage of initial mass remaining after two years for Metrosideros polymorpha leaf
litter originating in either the Woodland or the Woodland �
Grass site, decomposed in situ or in a common garden in the
seasonal submontane zone of Hawaii Volcanoes National
Park; n � 4 replicates.
** P �0.01.
1330
INVITED FEATURE
Ecological Applications
Vol. 11, No. 5
FIG. 4. Ecosystem nitrogen pools in a series of burned and unburned sites in the seasonal submontane zone of Hawaii
Volcanoes National Park. UB � unburned woodland, YB � young burn (sites burned in 1987), OB � old burn (sites burned
in 1970), and TB � twice burned (sites burned in 1970 and 1987). Data are means (n � 5 individuals for trees, n � 16
cores for soils, n � 8 plots for other ecosystem compartments). Within a compartment, bars with the same lowercase letter
above are statistically indistinguishable at P � 0.05 using a Tukey multiple-comparisons test.
cies dominance from Schizachyrium condensatum to
Melinis minutiflora.
Area-weighted decomposition of litter in the grassland was lower than in the unburned woodland (Fig.
1B), primarily due to the small pool of litter. Annual
net N mineralization from SOM was significantly higher in the Grassland than in either woodland treatment
(Table 1).
Aboveground production (Fig. 1C) and N uptake
(Table 1) in the Grassland were less than half of that
in the Woodland � Grass. Again, grasses produced a
similar amount of biomass and required a similar
amount of N whether they were alone or with natives
and they were unable to compensate for the production
and uptake of native species (Fig. 1C and Table 1).
Belowground production in the Grassland was 194
g·m�2·yr�1, less than half of the lower estimates for the
unburned treatments. Although we cannot statistically
compare root production between habitats because of
sampling methodology, it is clear that root production
in the unburned-woodland environments was substantially higher than in the Grassland. Root N uptake was
also correspondingly lower in the Grassland than the
unburned woodland, contributing to lower total uptake
values there (Table 1). Uptake appears to be uncoupled
from N availability in the absence of native species,
leading to a functionally different internal N cycle following fire. Increased net N mineralization combined
with reduced plant uptake increases the potential for
N loss from this site through leaching or trace-gas emissions (Likens et al. 1970, Vitousek and Matson 1984).
This pattern of increased soil N mineralization rates
with conversion to grassland is opposite to the pattern
seen in many continental tropical forest–grassland conversions where N mineralization rates are reduced relative to native forest (Neill et al. 1995, 1997, Johnson
and Wedin 1997). This difference may be partially explained by N loss during fire. In the conversion of
mature tropical forest to grassland, Kauffman et al.
(1998) reported losses of 41–66% of ecosystem N
pools. By contrast, we saw no significant change in
total ecosystem N pools with fire (Fig. 4). Differences
may also be due to species differences in litter quality.
In our sites, grass litter was somewhat more decomposable than native litter, which led to higher mineralization from SOM (Table 1). In dry shrublands where
M. minutiflora invaded in the absence of fire, Asner
and Beatty (1996) found a similar pattern of higher N
availability under the grass than under native species.
Effect of fire on ecosystem N pools
Nitrogen in the aboveground compartments differed
significantly among burned and unburned sites (Fig.
4). This difference was primarily due to the loss of
trees, which comprised the largest aboveground N pool
in the unburned site. The dominant tree species across
this range of sites, Metrosideros polymorpha, is killed
by intense fires (Tunison et al. 1995), and at our sites
�25% of trees at the young burn site were still living.
Dead trees and woody biomass represent about half of
the N pools in aboveground biomass after one fire, a
pattern that is consistent with other studies of forest to
grassland conversion (Kauffman et al. 1995, 1998).
However, woody biomass declined with successive
fires and was a very small part of total ecosystem N
pools at the twice-burned sites. Fire also eliminated the
October 2001
INVASIVE SPECIES AND THE SOIL
1331
FIG. 5. Loss of nitrogen from above- and belowground pools during controlled burning of grass-invaded sites in the
submontane zone of Hawaii Volcanoes National Park. (A) Low-intensity burn, January 1993; (B) High-intensity burn, June
1995. Measurements of standing biomass and N pools were taken within two weeks prior to and immediately after fire. Data
are means. Significant differences between before and after values using a students t test: *P � 0.05, **P � 0.01; NS � not
significantly different.
dominant understory shrubs Styphelia tameiameia and
Osteomeles anthyllidifolia (Hughes et al. 1991). Since
no species completely replaces these, biomass N in
burned areas is lower than in the unburned sites. Biomass N in grasses is not significantly different between
burned and unburned sites (C. M. D’Antonio, unpublished data).
The single largest pool of nitrogen in these ecosystems is in the soil (Fig. 4). Soil N pools were not
significantly different between burned and unburned
sites in our site survey suggesting either that past fires
were not intense enough to cause combustion losses of
N from soil, or that fires are still too infrequent to have
removed N at a rate greater than inputs. Although N
combusted from soil may have been replaced by deposition of ash from combusted aboveground biomass
(Raison 1979, Dyrness et al. 1986), this would only
account for an addition of �1% of total soil N (Fig.
4).
Direct losses of N with fire were larger in our highintensity fire than in our low-intensity burn (Fig. 5).
Aboveground biomass was reduced by 90% in the highintensity fire but only by 48% in the low-intensity burn.
This resulted in similar relative losses of aboveground
N (Fig. 5). As with our site survey, soil organic matter
was the single largest pool of ecosystem N. It was not
affected by fire in our low-intensity burn but was reduced by 20% in our high-intensity fire (Fig. 5). This
large loss of N from SOM suggests that, in this system,
fire severity impacts N loss primarily though the
amount of SOM that is consumed. This pattern is similar to that seen in other systems where the majority
of ecosystem N is contained in litter and soil (Dyrness
et al. 1986), which is in direct contrast to systems where
Ecological Applications
Vol. 11, No. 5
INVITED FEATURE
1332
TABLE 2. Asymbiotic nitrogen fixation in association with various substrates in the unburned woodland and sites that had
burned twice and are now grassland.
Site
Substrate
Unburned
Twice burned
Metrosideros polymorpha
Leaf litter
Woody debris
Soil organic layer
Dodonaea viscosa
Leaf litter
Total
M. polymorpha
Woody debris
D. viscosa
Leaf litter
Total
Biomass
(g/m2)
C2H4
production
(nmol·
g�1·h�1)
Conversion
factor
(C2H4:N2)
Total
contribution
(kg N·
ha�1·yr�1)
170
1377
600
1.50
0.53
0.30
3.45
10.16
1.34
0.152
0.146
0.271
15
8.31
12.10
0.021
0.590
313
0.28
1.07
0.165
15
6.65
13.37
0.015
0.180
Note: Values for acetylene reduction rates and conversion factors are from Ley and D’Antonio (1998).
ecosystem N is more evenly partitioned between soil
and plant biomass (e.g., Kauffman et al. 1995), and
biomass combustion determines N loss. The lack of
difference in soil N pools among our survey sites was
most likely due to the low intensity of fires in these
sites.
Effect of fire on asymbiotic N fixation
Ley and D’Antonio (1998) found that only two substrates showed N fixation activity in repeatedly burned
sites. These were Dodonaea viscosa leaf litter and
woody debris from Metrosideros polymorpha. Despite
repeated sampling of grass roots, rhizosphere soil, and
grass litter, we found no evidence of N fixation in association with grasses. Unlike other grass-dominated
systems where N fixation has been observed to increase
after fire (e.g., Eisele et al. 1989), observations from
our controlled burns suggest that there is no fixation
associated with grasses or soil after fire (R. Ley and
C. M. D’Antonio, unpublished data). This is also in
contrast to several studies of tropical grasses, which
suggest that they often have fixation associated with
their root zone (DePolli et al. 1977, Robertson and
Rosswall 1986, Abbadie et al. 1992, Boddey and Dobereiner 1995). These studies suggest that fixation occurs in environments with low partial pressure of oxygen perhaps associated with wet tropical soils. The
lack of fixation associated with grasses in our study
may be due to the relatively well-drained nature of the
soils in our sites or the lack of suitable microbes.
Metrosideros polymorpha leaf litter was a major
source of asymbiotic N fixation in the unburned woodland as was microbial activity in the O-layer of the soil
(Table 2). As a consequence of fire-driven conversion
from woodland to grassland, the capacity of the system
to accrete N by biological fixation appears to be di-
minished. Our annual estimates of N inputs via asymbiotic fixation for the unburned woodland were many
times higher than those from the twice-burned sites
(Table 2). This difference is due largely to the loss of
the native tree M. polymorpha and a change in the Olayer: we did not find fixation associated with the Olayer in the twice-burned site. However, we did find
asymbiotic fixation in association with woody debris
in both burned and unburned sites. Thus after the initial
fire that kills M. polymorpha, it is possible that N fixation inputs might increase because of the enormous
amount of dead wood that gradually falls to the forest
floor. Likewise, Wei and Kimmins (1998) found increases in asymbiotic N fixation after fire in a lodgepole
pine forest due to increases in woody debris on the
forest floor. In our sites most of this woody material
is volatilized in a second fire, and since no trees regenerate between fires (Hughes and Vitousek 1991)
there are no new inputs of woody material to the soil
surface. This results in reduced woody substrates for
fixation in repeatedly burned areas. Due to the low
levels of asymbiotic fixation in the absence of nativelitter substrates and low levels of atmospheric N deposition (Vitousek et al. 1987, Heath and Huebert
1999), it is unlikely that the N lost through combustion
of aboveground biomass during fire will be replaced
prior to the next fire, as has been found in studies of
ecosystems where fire is a natural disturbance (Kauffman et al. 1993, Ojima et al. 1994).
Conclusions
Our results in the unburned woodland support the
predictions of Chapin et al. (1996): when an invader
differs from natives only in traits related to the focal
ecosystem process (in this case N cycling between
plants and soil) that are continuously distributed, and
October 2001
INVASIVE SPECIES AND THE SOIL
not discrete, changes will occur slowly and will be
difficult to detect. The changes that we saw with grass
invasion into the unburned woodland were related to
competitive interactions (reduction of native litter production) and to higher decomposability of grass litter
compared to native litter and did not translate to an
alteration of N cycling between plants and soil. The
largest and most rapid changes that we saw in ecosystem N dynamics, including inputs, losses, and internal
N cycling, were due to the direct effects of fire and the
loss of native species with fire. Thus, it is the suite of
grass traits that affect fuel loading in the ecosystem
that had the greatest ultimate consequences for ecosystem N dynamics because these traits ultimately lead
to the removal of native species from the system. In
the Grassland, exotic grasses could not compensate for
the loss of native species in terms of N accretion via
fixation substrates, N uptake, and N contributions in
litterfall. We suggest that the mechanistic, trait-based
approach is a useful framework for predicting the impacts of invaders on ecosystem processes. However, it
requires a priori knowledge of traits of both the invader
and the invaded system in relation to target ecosystem
processes and anticipation of important ecosystem
functions that may be altered by invaders. Ecologists
will therefore need to work across levels of organization and integrate their research with land managers
to sharpen our ability to understand and prioritize the
effects of exotic species.
ACKNOWLEDGMENTS
We would like to thank Peter Vitousek for providing laboratory facilities and logistical support, and David Foote,
Charles Stone, and Tim Tunison of Hawaii Volcanoes National Park for providing logistical support and site access
and J. Manassian and the HVNP fire Crew for conducting the
experimental burns. Numerous assistants, volunteer and paid,
helped with fieldwork. Special thanks go to Xinghua Zeng,
Heather Swartz, Nichole Barger, and Susan Cordell and to
volunteers associated with the UREP program at University
of California-Berkeley. Financial support was provided by a
NASA Global Change Graduate Fellowship, an NSF Dissertation Improvement Grant, and a U.C. Vice Chancellor’s Research Grant to M. C. Mack, and by NSF BSR 9119618 and
UCB Committee on Research funds to C. M. D’Antonio.
LITERATURE CITED
Abbadie, L., A. Mariotti, and J. C. Menaut. 1992. Independence of savanna grasses from soil organic matter for their
nitrogen supply. Ecology 73:608–613.
Asner, G. P., and S. W. Beatty. 1996. Effects of an African
grass invasion on Hawaiian shrubland nitrogen biogeochemistry. Plant and Soil 186:205–211.
Berendse, F., R. Bobbink, and G. Rouwenhorst. 1989. A comparative study on nutrient cycling in wet heathland ecosystems. II. Litter decomposition and nutrient mineralization. Oecologia 78:338–348.
Berendse, F., H. Oudhof, and J. Bol. 1987. A comparative
study on nutrient cycling in wet heathland ecosystems. I.
Litter production and nutrient losses from the plant. Oecologia 74:174–184.
Boddey, R. M., and U. Dobereiner. 1995. Nitrogen fixation
associated with grasses and cereals: recent progress and
perspectives for the future. Fertilizer Research 42:241–250.
1333
Boring, L. R., and W. T. Swank. 1984. The role of black
locust (Robinia pseudoacacia) in forest succession. Journal
of Ecology 72:749–766.
Breshears, D. D., J. W. Nyhan, C. E. Heil, and B. P. Wilcox.
1998. Effects of woody plants on microclimate in a semiarid woodland: soil temperature and evaporation in canopy
and intercanopy patches. International Journal of Plant Sciences 159:1010–1017.
Caldwell, M. M., and R. A. Virginia. 1989. Root systems.
Pages 367–398 in R. W. Pearcy, J. Ehleringer, H. A. Mooney, and P. W. Rundel, editors. Plant physiological ecology:
field methods and instrumentation. Chapman & Hall, New
York, New York, USA.
Chapin, F. S., III, H. Reynolds, C. M. D’Antonio, and V.
Eckhart. 1996. The functional role of species in terrestrial
ecosystems. Pages 403–430 in B. Walker and W. Steffan,
editors. Global change in terrestrial ecosystems. Cambridge
University Press, Cambridge, UK.
Christian, J. M., and S. D. Wilson. 1999. Long-term impacts
of an introduced grass in the northern Great Plains. Ecology
80:2397–2407.
Crews, T. E., K. Kitayama, J. H. Fownes, R. H. Riley, D. A.
Herbert, D. Mueller-Dombois, and P. M. Vitousek. 1995.
Changes in soil phosphorus fractions and ecosystem dynamics across a long chronosequence in Hawaii. Ecology
76:1407–1424.
D’Antonio, C. M., T. L. Dudley, and M. C. Mack. 1999.
Disturbance and biological invasions: direct effects and
feedbacks. Pages 429–468 in L. R. Walker, editor. Ecosystems of disturbed ground. Elsevier, Amsterdam, The
Netherlands.
D’Antonio, C. M., R. F. Hughes, M. Mack, D. Hitchcock,
and P. M. Vitousek. 1998. The response of native species
to removal of invasive exotic grasses in a seasonally dry
Hawaiian woodland. Journal of Vegetation Science 9:699–
712.
D’Antonio, C. M., J. T. Tunison, and R. K. Loh. 2000. Variation in impact of exotic grasses on native plant composition in relation to fire across an elevation gradient in
Hawai’i. Austral Ecology 25:507–522.
D’Antonio, C. M., and P. M. Vitousek. 1992. Biological invasion by exotic grasses, the grass/fire cycle and global
change. Annual Review of Ecology and Systematics 23:
63–87.
DePolli, H., E. Matsui, and J. Dobereiner. 1977. Confirmation
of nitrogen fixation in two tropical grasses by 15N2 incorporation. Soil biology and Biochemistry 9:119–123.
Dyrness, C. T., L. A. Viereck, and K. Van Cleve. 1986. Fire
in taiga communities of interior Alaska. Pages 75–85 in K.
Van Cleve, F. S. Chapin III, P. W. Flanagan, L. A. Viereck,
and C. T. Dyrness, editors. Forest ecosystems in the Alaskan taiga. Springer-Verlag, New York, New York, USA.
Eisele, K. A., D. S. Schimel, L. A. Kapustka, and W. J. Parton.
1989. Effects of available P and N:P ratios on non-symbiotic dinitrogen fixation in tallgrass prairie soils. Oecologia 79:471–474.
Fisher, M. M., I. M. Rao, M. A. Ayarza, C. E. Lascano, J. I.
Sanz, R. J. Thomas, and R. R. Vera. 1994. Carbon storage
by introduced deep-rooted grasses in the South American
savannas. Nature 371:235–238.
Hart, S. C., and M. K. Firestone. 1989. Evaluation of three
in situ soil nitrogen availability assays. Canadian Journal
of Forest Research 19:185–191.
Heath, J. A., and B. J. Huebert. 1999. Cloudwater deposition
is a source of fixed nitrogen in a Hawaiian montane forest.
Biogeochemistry 44:119–134.
Holcomb, R. T. 1987. Eruptive history and longterm behaviour of Kilauea volcano. Pages 261–350 in R. W. Decker,
T. L. Wright, and P. H. Stauffer, editors. Volcanism in Ha-
1334
INVITED FEATURE
wai’i. Professional Paper 1350. U.S. Geological Survey,
Menlo Park, California, USA.
Holmes, T. H., and K. J. Rice. 1996. Patterns of growth and
soil-water utilization in some exotic annuals and native
perennial bunchgrasses of California. Annals of Botany 78:
233–243.
Hooper, D. U., and P. M. Vitousek. 1998. Effects of plant
composition and diversity on nutrient cycling. Ecological
Monographs 68:121–149.
Hughes, F., and P. M. Vitousek. 1993. Barriers to the reestablishment of native shrubs following fire in the seasonal
submontane zone of Hawai’i. Oecologia 557–563.
Hughes, R. F., P. M. Vitousek, and J. T. Tunison. 1991. Alien
grass invasion and fire in the seasonal submontane zone of
Hawai’i. Ecology 72:743–746.
Joffre, R. 1990. Plant and soil nitrogen dynamics in Mediterranean grasslands: a comparison of annual and perennial
grasses. Oecologia 85:142–149.
Johnson, N. C., and D. A. Wedin. 1997. Soil carbon, nutrients, and mycorrhizae during conversion of dry tropical
forest to grassland. Ecological Applications 7:171–182.
Kauffman, J. B., D. L. Cummings, and D. E. Ward. 1998.
Fire in the Brazilian Amazon. 2. Biomass, nutrient pools
and losses in cattle pastures. Oecologia 113:415–427.
Kauffman, J. B., D. L. Cummings, D. E. Ward, and R. Babbitt.
1995. Fire in the Brazilian Amazon. 1. Biomass, nutrient
pools and losses in slashed primary forests. Oecologia 104:
397–408.
Kauffman, J. B., D. L. Sanford, D. L. Cummings, I. H. Salcedo, and E. V. S. B. Sampaio. 1993. Biomass and nutrient
dynamics associated with slash pine fires in neotropical dry
forests. Ecology 74:140–151.
Kochy, M., and S. D. Wilson. 1997. Litter decomposition
and nitrogen dynamics in Aspen forest and mixed-grass
prairie. Ecology 78:732–739.
Kummerow, J., J. Castillanos, M. Maas, and A. Larigauderie.
1990. Production of fine roots and the seasonality of their
growth in a Mexican deciduous dry forest. Vegetatio 90:
73–80.
Le Maitre, D. C., B. W. Van Wilgen, R. A. Chapman, and D.
H. McKelly. 1996. Invasive plants and water resources in
the Western Cape Province, South Africa: modeling the
consequences of a lack of management. Journal of Applied
Ecology 33:161–172.
Ley, R., and C. M. D’Antonio. 1998. Exotic grass invasion
alters potential rates of N fixation in Hawaiian woodlands.
Oecologia 113:179–187.
Likens, G. E., F. H. Bormann, N. M. Johnson, D. W. Fisher,
and R. S. Pierce. 1970. Effects of forest cutting and herbicide treatment on nutrient budgets in the Hubbard Brook
watershed-ecosystem. Ecological Monographs 40:23–47.
Mack, M. C. 1998. Effect of exotic grass invasion on ecosystem nitrogen dynamics in a Hawaiian woodland. Dissertation. University of California, Berkeley, California,
USA.
Mack, M. C., and C. M. D’Antonio. 1998. Impacts of biological invasions on disturbance regimes. Trends in Ecology and Evolution 13:195–198.
Maron, J. L., and R. L. Jeffries. 1999. Bush lupine mortality,
altered resource availability, and alternative vegetation
states. Ecology 80:443–454.
Musil, C. F., and G. F. Midgley. 1990. The relative impact
of invasive Australian Acacias, fire and season on the soil
chemical status of a sand plane lowland fynbos community.
South African Journal of Botany 56:417–419.
Nadelhoffer, K. J., and J. W. Raich. 1992. Fine root production estimates and belowground carbon allocation in forest
ecosystems. Ecology 73:1139–1147.
Neill, C., M. C. Piccolo, C. C. Cerri, P. A. Steudler, J. M.
Melillo, and M. Brito. 1997. Net nitrogen mineralization
Ecological Applications
Vol. 11, No. 5
and net nitrification rates in soils following deforestation
for pasture across the southwestern Brazilian Amazon Basin landscape. Oecologia 110:243–252.
Neill, C., M. C. Piccolo, P. A. Steudler, J. M. Melillo, B. J.
Feigl, and C. C. Cerri. 1995. Nitrogen dynamics in soils
of forests and active pastures in the western Brazilian Amazon Basin. Soil Biology and Biochemistry 27:1167–1175.
Ojima, D. S., D. S. Schimel, W. J. Parton, and C. E. Owensby.
1994. Long- and short-term effects of fire on nitrogen cycling in tallgrass prairie. Biogeochemistry 24:67–84.
Paul, E. A., and F. E. Clark. 1989. Soil microbiology and
biochemistry. Academic Press, San Diego, California,
USA.
Raison, R. J. 1979. Modifications of the soil environment by
vegetation fires, with particular reference to nitrogen transformations: a review. Plant and Soil 51:73–108.
Raison, R. J., P. K. Khanna, and P. V. Woods. 1985. Mechanisms of element transfer to the atmosphere during vegetation fires. Canadian Journal of Forest Research 15:132–
140.
Robertson, G. P., and T. Rosswall. 1986. Nitrogen in West
Africa: the regional cycle. Ecological Monographs 56:43–
72.
Schlesinger, W. H., and A. M. Pilmanis. 1998. Plant–soil
interactions in deserts. Biogeochemistry 42:169–187.
Scott, N. A., and D. Binkley. 1997. Foliage litter quality and
annual net N mineralization: comparison across North
American forest sites. Oecologia 111:151–159.
Smith, C. W., and J. T. Tunison. 1992. Fire and alien plants
in Hawai’i: research and management implications for native ecosystems. Pages 394–408 in C. P. Stone, T. Tunison,
and D. A. Stone, editors. Alien plant invasions in native
ecosystems of Hawai’i: management and research. University of Hawaii Press, Honolulu, Hawai’i, USA.
Stock, W. D., K. T. Wienand, and A. C. Baker. 1995. Impacts
of invading N2-fixing Acacia species on patterns of nutrient
cycling in two Cape ecosystems: evidence from soil incubation studies and 15N natural abundance values. Oecologia 101:375–382.
Swift, M. J., O. W. Heal, and J. M. Anderson. 1979. Decomposition in terrestrial ecosystems. Blackwell Scientific, Oxford, UK.
SYSTAT. 1997. Systat for Windows: statistics, version 7
Edition. SYSTAT, Evanston, Illinois, USA.
Tilman, D., D. Wedin, and J. Knops. 1996. Productivity and
sustainability influenced by biodiversity in grassland ecosystems. Nature 379:718–720.
Torn, M., S. E. Trumbore, O. A. Chadwick, P. M. Vitousek,
and D. M. Hetricks. 1997. Mineral control of soil organic
carbon storage and turnover. Nature 389:170–173.
Tunison, J. T., R. L. Loh, and J. A. Leialoha. 1995. Fire
effects in the submontane seasonal zone Hawai’i Volcanoes
National Park. Cooperative Agreement CA 8007-2-9004,
University of Hawaii—Manoa. Cooperative National Park
Resources Study Unit Technical Report 97.
Van Wilgen, B. W., R. M. Cowling, and C. J. Burgers. 1996.
Valuation of ecosystem services: a case study from South
African fynbos ecosystems. BioScience 46:184–189.
Vitousek, P. M. 1982. Nutrient cycling and nutrient-use efficiency. American Naturalist 119:553–572.
Vitousek, P. M., G. Gerrish, D. Turner, L. Walker, and D.
Mueller-Dombois. 1995. Litterfall and nutrient cycling in
four Hawaiian montane rainforests. Journal of Tropical
Ecology 11:189–203.
Vitousek, P. M., and P. A. Matson. 1984. Mechanisms of
nitrogen retention in forest ecosystems: a field experiment.
Science 225:51–52.
Vitousek, P. M., and L. R. Walker. 1989. Biological invasion
by Myrica faya in Hawai’i: plant demography, nitrogen
October 2001
INVASIVE SPECIES AND THE SOIL
fixation, ecosystem effects. Ecological Monographs 59:
247–265.
Vitousek, P. M., L. R. Walker, L. D. Whiteaker, D. MuellerDombois, and P. M. Matson. 1987. Biological invasion by
Myrica faya alters ecosystem development in Hawaii. Science 238:802–804.
Vogt, K. A., and H. Persson. 1991. Measuring growth and
development of roots. Pages 777–501 in J. P. Lassoie and
T. M. Hinckley, editors. Techniques and approaches in forest tree physiology. CRC Press, Boca Raton, Florida, USA.
Wedin, D. A., and D. Tilman. 1990. Species effects on ni-
1335
trogen cycling: a test with perennial grasses. Oecologia 84:
433–441.
Wei, X., and J. P. Kimmins. 1998. Asymbiotic nitrogen fixation in harvested and wildfire-killed lodgepole pine forests
in the central interior of British Columbia. Forest Ecology
and Management 10:343–353.
Witkowski, E. T. F. 1991. Effects of invasive alien Acacias
on nutrient cycling in the coastal lowlands of the Cape
fynbos. Journal of Applied Ecology 28:1–15.
Wolf, E. W., and J. Morris. 1996. Geological map of the
island of Hawaii. United States Geological Survey, Menlo
Park, California, USA.