Stable nitrogen isotope patterns of trees and soils altered
by long-term nitrogen and phosphorus addition to a lowland
tropical rainforest
Jordan R. Mayor • S. Joseph Wright •
Edward A. G. Schuur • Mollie E. Brooks •
Benjamin L. Turner
Received: 14 July 2013 / Accepted: 6 February 2014 / Published online: 1 March 2014
 Springer International Publishing Switzerland 2014
Abstract Foliar nitrogen (N) isotope ratios (d15N) are
used as a proxy for N-cycling processes, including the
‘‘openness’’ of the N cycle and the use of distinct N
sources, but there is little experimental support for such
proxies in lowland tropical forest. To address this, we
examined the d15N values of soluble soil N and canopy
foliage of four tree species after 13 years of factorial N
and P addition to a mature lowland rainforest. We
hypothesized that N addition would lead to 15N-
enriched soil N forms due to fractionating losses,
whereas P addition would reduce N losses as the plants
and microbes adjusted their stoichiometric demands.
Chronic N addition increased the concentration and
d15N value of soil nitrate and d15N in live and senesced
leaves in two of four tree species, but did not affect
ammonium or dissolved organic N. Phosphorus
addition significantly increased foliar d15N in one tree
species and elicited significant N 9 P interactions in
two others due to a reduction in foliar d15N enrichment
under N and P co-addition. Isotope mixing models
indicated that three of four tree species increased their
use of nitrate relative to ammonium following N
addition, supporting the expectation that tropical trees
use the most available form of mineral N. Previous
observations that anthropogenic N deposition in this
tropical region have led to increasing foliar d15N values
over decadal time-scales is now mechanistically linked
to greater usage of 15N-enriched nitrate.
Keywords Ecosystem ecology  Gigante
Fertilization Experiment  Mass balance mixing
models  Panama  Stoichiometry
Introduction
Stable N isotope ratios (15N:14N expressed as d15N)
provide a non-destructive tracer that integrates plant
Responsible Editor: Jonathan Sanderman.
Electronic supplementary material The online version of
this article (doi:10.1007/s10533-014-9966-1) contains supple-
mentary material, which is available to authorized users.
J. R. Mayor (&)  S. J. Wright  B. L. Turner
Smithsonian Tropical Research Institute, Apartado
0843-03092, Balboa, Ancon, Republic of Panama
e-mail: clavulina@gmail.com
Present Address:
J. R. Mayor
Department of Forest Ecology & Management, Swedish
University of Agricultural Sciences, 90183 Umea˚,
Sweden
E. A. G. Schuur  M. E. Brooks
Department of Biology, University of Florida,
Gainesville, FL 32611, USA
Present Address:
M. E. Brooks
Institute of Evolutionary Biology and Environmental
Studies, University of Zurich, Zurich, Switzerland
123
Biogeochemistry (2014) 119:293–306
DOI 10.1007/s10533-014-9966-1
responses to their abiotic and biotic environments
through time (Dawson et al. 2002). Several studies
have demonstrated a relationship between ecosystem
d15N values and the ‘‘openness’’ of the N cycle (e.g. a
ratio of production and demand to loss). For instance,
d15N measurements have been used to trace pathways
of forest N losses in gaseous form (N2O and N2) or in
leachate [NO3
- and dissolved organic N (DON)] at
both local (Houlton et al. 2006; Brookshire et al. 2012;
Koehler et al. 2009) and regional scales (Bai and
Houlton 2009). High rates of N loss influence d15N in
the remaining soil pools in part through the strongly
fractionating process of denitrification (Pe´rez et al.
2000) which preferentially consumes 14N and can yield
relatively 15N-enriched soil nitrate remaining in soil
solutions (reviewed in Hobbie and Ouimette 2009).
This is a main mechanism whereby a ‘‘loosening’’ of
the N cycle leads to enriched soil and plant d15N values.
The ability of foliar d15N to act as an N cycling proxy
has also enabled detection of regional increases in N
availability due to agricultural intensification (Park
et al. 2011), succession (Wang et al. 2007; Hobbie et al.
2005), afforestation (Davidson et al. 2007), and
atmospheric deposition (Hietz et al. 2011).
Others studies have shown that plant d15N values
trace uptake of specific N forms, even in N-rich
tropical rainforests. For example, plant d15N values
reflected a community-wide switch between two
mineral N sources along a precipitation gradient in
Hawaii (Houlton et al. 2006; Schuur and Matson 2001)
and among spatially distinct N sources in litter and
mineral soils of French Guiana (Schimann et al. 2008).
Experimental additions of 15N-enriched tracers have
shown that plants prefer mineral over organic forms of
N, even in N limited ecosystems (Harrison et al. 2007;
Ashton et al. 2008), and that local adaptations to
growth conditions in part determine N uptake rates and
preferences for ammonium versus nitrate (Wang and
Macko 2011; Andersen and Turner 2013).
The complexity of the tropical forest N cycle, with
high rates of biological activity, adds to the difficulty in
assessing the influence of perturbations (Hedin et al.
2009) and isolating the dominant controls on plant and
soil d15N values (Pardo and Nadelhoffer 2010). Foliar and
total soil d15N values show strong latitudinal gradients
(Martinelli et al. 1999; Amundson et al. 2003), promoted
by enhanced loss of isotopically light forms of N (Pe´rez
et al. 2006) and more rapid microbial processing at lower
latitudes (Kramer et al. 2003; Dijkstra et al. 2008). Once
thought common to industrialized regions alone, isotopic
signals of atmospheric N deposition have now been
detected in plant tissue from tropical regions as well
(Hietz et al. 2010, 2011). However, experimental
evidence in support of the use of plant, soil, or (collec-
tively) ecosystem d15N values as indicators of a changing
N cycle in tropical forests remains scarce. In particular, no
study has experimentally examined the link between N
and P availability and thed15N values of extractable N and
plants in a mature lowland tropical forest.
We used the longest running fertilization experiment in
lowland tropical forest to test hypotheses linking nutrient
availability to foliar d15N. We measured d15N values in
soluble soil N pools, leaves, and litterfall, beneath four
locally common tree species following 13 years of
factorial N and P fertilization. Previous studies in the
experiment showed complexity in growth responses of the
plant community: P addition increased fine litterfall,
potassium addition reduced fine root growth and influ-
enced seedling growth, and N addition in conjunction with
potassium ameliorated otherwise stand-wide trends
towards slower relative growth rates (Wright et al. 2011;
Santiago et al. 2012). Further, species-specific responses
in leaf N and P accumulation were observed among four
common tree species (Mayor et al. 2014). We hypothe-
sized that chronic N addition would enrich soluble d15N
values due to greater N losses, and that this signal would
be seen in the foliage of all tree species examined. In
contrast, we hypothesized that P addition would lead to
depletion of plant d15N values due to a tightening of the N
cycle as plants and microbes sought to rebalance stoichi-
ometric demands of growth and metabolism (Niklas et al.
2005). We specifically examined: (1) whether chronic N
and P addition affected d15N values of extractable soil
nitrate, ammonium, and dissolved organic N; (2) whether
changes in d15N values of soil N pools correspond to d15N
values of canopy foliage; (3) whether N resorption from
litterfall and translocation within plants could contribute
to altered foliar d15N; and, (4) the extent to which altered
N and P availability influence the forms of N used by
tropical trees.
Methods
Site description and experimental design
The Gigante Fertilization Experiment occupies a
38.4 ha plot of mature ([200 years) tropical moist
294 Biogeochemistry (2014) 119:293–306
123
forest on the Gigante peninsula in the Barro Colorado
Natural Monument, Panama. Elevation is 25–61 m
a.s.l., mean annual temperature is 27 ± 0.1 C, and
mean annual rainfall is 2,650 ± 146 mm (1995–2007)
with a pronounced 4 month dry season from Decem-
ber to April (Wright et al. 2011). The site is underlain
by Miocene basalt with soils classified as Oxisols and
Inceptisols on higher and lower parts of the landscape,
respectively, with clay contents of *70 % (Turner
et al. 2013). Soil extractions from control plots
indicated very low concentrations of readily available
phosphate (\1 mg P kg-1) determined from Mehlich-
3 solutions or anion-exchange membrane extractions
(Turner et al. 2013).
Beginning in June 1998 a factorial NPK and a
second micronutrient fertilization experiment were
established in thirty-six 40 9 40 m plots. All treat-
ments were replicated once in each of four blocks
along the elevation gradient to control for spatial
variation in soil properties (Yavitt et al. 2009).
Fertilizer applications occurred four times a year,
evenly spaced within the rainy season (May to
November; Supplementary Fig. S1). The current study
sampled from sixteen plots in a factorial N 9 P design
(i.e. four plots each of control, ?N, ?P, and ?N ? P),
with total annual applications of 125 kg N ha-1 year-1
as urea ((NH2)2CO) and 50 kg P ha
-1 year-1 as triple
superphosphate (Ca(H2PO4)2H2O). Annual rates of N
addition were comparable to the annual inputs from
litterfall (Yavitt et al. 2011). The d15N value of the
urea fertilizer was -2.2 %.
Plant sampling
Four tree species were selected from within the Gigante
Fertilization Experiment because at least two mature
individuals were present in 95 % of the 16 9 4 species-
by-plot combinations (J. Wright, unpublished data;
Supplementary Table S1). Sun exposed canopy leaves
of the medium sized canopy tree Alseis blackiana
(Rubiaceae), the medium sized sub-canopy tree Heis-
teria concinna (Olacaceae), and the large canopy tree
Tetragastris panamensis (Burseraceae) were obtained
in March 2011 (i.e. mid dry season) using steel
ammunition. All four species associate with arbuscular
mycorrhizal fungi (M. Sheldrake, unpublished data),
thus accounting of ectomycorrhizal fractionation is not
necessary (Hobbie and Ho¨gberg 2012).
Freshly senesced litter was collected once every
2 weeks from 1 m2 mesh litterfall traps located
beneath target species between November 2010 and
March 2011 (i.e. late wet season to mid dry season).
Fresh and senesced fronds of a medium sized, clonal,
sub-canopy palm, Oenocarpus mapora (Arecaceae)
were obtained in April 2011 with a pole cutter. Leaflets
of each O. mapora were taken from a minimum of nine
fronds originating from three separate shoots and
composited by tree (fresh) or plot (senesced). Leaf
litterfall N concentration and d15N values were mea-
sured to assess whether N resorption during leaf
senescence alters litter d15N values.
All leaf tissue (excluding petioles) was cleaned of
epiphylls and damaged tissue before being dried at
60 C for 48 h, and ground to a fine powder. Isotope
analyses were performed using a Flash 1112 Series
Elemental Analyzer coupled through a Conflo III
interface to a Delta V Advantage continuous flow
isotope ratio mass spectrometer (Thermo Scientific,
Bremen, Germany) at the Smithsonian Tropical
Research Institute‘s Stable Isotope Laboratory. Stable
isotope abundances are reported as: d15N = (Rsample/
Rstandard - 1) 9 1000, where R =
15N:14N ratio of the
sample or standard. Based on three internal standards,
run error rates were 0.05 % and analytical precision for
leaf samples analyzed in triplicate had an average
standard error of 0.06 %.
Soil nutrient sampling
Ion exchange resin bag incubations were used to
compare in situ accumulation rates of soluble nitrate,
ammonium, and phosphate (Giblin et al. 1994; Schuur
and Matson 2001) and to provide a time-integrated
source of d15N values from mineral N. During three
time periods (March 24–April 21, June 15–July 16, and
September 22–October 19; hereafter referred to as
early, mid, and late rainy season; Supplementary Fig.
S1), eight mesh bags (220 lm mesh, 40 cm2) contain-
ing either 3 g of anion or cation exchange resins
(Biorad, AG1/50W-X8, #140/142-1421) were
inserted at a 45 angle 5–10 cm deep at four random
locations in each of the 16 plots. Due to logistical
constraints associated with large sample sizes, d15N
values for both NO3
- and NH4
? were fully replicated
from the mid rainy season resin bag incubations,
partially replicated (half the treatment plots) for NO3
-
in the early rainy season, and fully replicated for NO3
-
Biogeochemistry (2014) 119:293–306 295
123
only in the late rainy season. The rainy season is a
period of high biological activity, associated with leaf
flushes (J. Wright, pers. obsv.) and increases in
microbial N and P (Turner and Wright 2014). After
removal from the field, resin bags were rinsed with
deionized water to remove adhering soil, extracted with
100 mL of weakly-acidified NaCl (Giblin et al. 1994),
filtered through Whatman GF/A filters, and immedi-
ately frozen at -36 C prior to colorimetric and
isotopic analyses.
Five soil cores were collected to 10 cm depth from
each plot in January 2011 using a 2.5 cm diameter soil
corer. Within 8 h of collection, roots, stones, and
mesofauna were removed and subsamples of each
homogenized core (*10 g fresh soil) were extracted
in 50 mL of 2 M KCl on a shaker table for 1 h, filtered
(GF/A), and immediately frozen at -36 C until
colorimetric determination of nitrate ? nitrite
(NO3
- ? NO2
-), ammonium (NH4
?), or phosphate
(PO4
-3) on a Lachat QuikChem 8500 (Hach Ltd.,
Loveland, CO, USA). Diluted (0.5 M salt) aliquots
(1.5 mL) of total dissolved N (TDN) and cation resin
ammonium extracts were converted to nitrate using
potassium peroxodisulfate thermo-oxidation (Cabrera
and Beare 1993; Doyle et al. 2004), and converted to
N2O gas for determination of d
15N and d18O values
using anaerobic cultures of Pseudomonas aureofasc-
iens denitrifying bacteria (Sigman et al. 2001; Knapp
et al. 2005). This yielded d15N values for TDN in KCl
extracts, d15N of ammonium in cation resin extracts,
and d15N and d18O values in both KCl and anion resin
nitrate extracts. Denitrified samples were analyzed
using a GC-Pal gas sampler coupled to a Thermo-
Finnigan continuous flow isotope ratio mass spec-
trometer at the University of Florida. Run error rates
for 15N and 18O were typically better than 0.2 % as
compared to NIST N-3 (National Institute of Stan-
dards and Technology) and USGS #32, 34, and 35
standards (IAEA 2004), respectively. To assure 100 %
oxidation efficiency of the NH4
? and TDN samples, a
range of standards (glycine, and amino caproic acid)
were oxidized with samples (concentrations of 1, 5,
and 10 mg N L-1, representing the range of sample
concentrations). Resulting reagent blanks derived
from purified (recrystallized three times) ACS-grade
peroxodisulfate (K2S2O8) solutions were removed
from each resulting d15N value on a per batch basis
(Knapp et al. 2005). The analytical precision of
oxidized samples duplicated between runs was on
average different by 1.29 % (n = 12) and 0.49 %
(n = 10) for NH4
? and TDN, respectively. This
analytical variability, although larger than unoxidized
standards, was deemed inconsequential to final plot
means because of independent replicate measure-
ments of each N form in each plot and the use of
standard deviations in mass balance estimations.
The concentration of DON was calculated as the
difference between TDN and mineral N
(nitrate ? ammonium) and d15NDON was calculated
using the following mass balance equation:
d15NDON ¼ d15NTDN  TDN½   d15NNH4

 NHþ4
 
þ d15NNO3  NO3½ Þ
 ð1Þ
where the N concentrations and the d15NTDN were
determined from 2 M KCl extractions and d15N values
of ammonium and nitrate were measured from fully-
replicated mid-rainy season resin bag incubations. On
average, DON pools were *12 times greater than
mineral N and were therefore relatively robust against
variability in mineral N d15N values. Resin accumu-
lated ion d15N values were preferred for this mass
balance calculation because of: their demonstrated
temporal stability (Templer and Weathers 2011); their
ability to concentrate ammonium in situ, rather than
using the high pH diffusion technique with associated
risks of fractionation during volatilization losses; and,
their month-long incubation eliminates some uncer-
tainty associated with spatially and temporally epi-
sodic mineral N production and loss processes
characteristic of tropical soils. It should be noted,
however, that resin and salt extractions may sample
slightly different pools of labile nitrate, partially as a
function of the anion exchange capacity of soils
(Lohse and Matson 2005), but this remains to be
determined.
Data analysis
Mixed effect models were fitted for leaf and soil
metrics using fixed treatments represented as dummy
variables of either ?N or ?P addition to assess
nutrient specific treatment effects and interactions. To
account for plot-level variance, Plot was fitted as a
random effect using the function lmer in the package
lme4 version 0.999999-0 (Bates et al. 2012) in R
version 2.15. We used Markov Chain Monte Carlo
sampling to test significance in each model using the
296 Biogeochemistry (2014) 119:293–306
123
mcmcsamp function because this is the least contro-
versial method for testing hypotheses in mixed effect
models (Baayan et al. 2008). Models were evaluated
by checking for symmetry and normality in posterior
distribution of parameters derived from 10,000 sam-
ples. Significance of fixed effects was determined with
p-values derived from ancillary functions calculated
from the posterior distribution of the parameters
(pMCMC) using the pvals.fnc function from the langu-
ageR package version 1.4 as these are conservative
with respect to sample sizes (Baayen 2011). For leaf
litterfall d15N and estimated proportional N use
derived from mixing models (discussed below), which
lack within-plot replication, linear models were fit
using the lm function in R with the four blocks
included as a fixed effect because fitting random
effects with fewer than five levels is not recommended
(Bolker et al. 2009).
To explore the relationship between d15N values in
plant tissue and available soil N, we used an isotope
mixing model approach similar to (Houlton et al.
2007). Initial underdetermined models assessed
whether foliar d15N values were interpretable as
mixtures of all three measured N sources (DON,
NH4
?, NO3
-) across all species and plots with or
without N fertilization using the optim package in R.
These initial underdetermined models fitted an uptake
parameter assuming source contributions were pro-
portional to soil N availability. Several measures of
soil N availability were used, including the 2 M KCl
and resin bag values reported here and monthly
measures taken after a decade of fertilization (Turner
et al. 2013). Despite the statistical power gained with
this approach, these general models were very poor
predictors (based on root mean square error) of plant
d15N values because leaves were on average 9 %
more depleted than DON. There was therefore no
reason to include DON as a potential source in this
system, as the modeled proportional contributions
would have been\0.01.
As a result of the elimination of DON as a potential
source, a better constrained two-source mixing model
termed IsoError (Phillips et al. 2005) was used on a
plot level instead. This Excel-based model propagates
plot-scale variances associated with plant and soil
measurements as it solves for the proportional fraction
(f ) of mineral N used by each species following mass
balance mixing models:
1 ¼ f NO3 þ f NH4 ð2Þ
d15Nfoliage ¼ d15NNO3  f NO3 þ d
15NNH4  f NH4
ð3Þ
d15Nfoliage ¼ d15Nfoliage  0:89 & ð4Þ
The *d15Nfoliage values were adjusted using fraction-
ation values obtained from paired leaf and excavated
fine root samples taken from seven other Panamanian
tree species. Averages from that study suggested a
regional enrichment factor of 0.89 % (n = 31 adult
trees from seven spp. on three soil types; unpublished
data) is more appropriate than the extra-tropical
2.0 % value used elsewhere (Houlton et al. 2007;
Handley et al. 1993). This fractionation value
accounts for the combined effects of internal plant N
allocation and the minor fractionation associated with
the transfer of N by arbuscular mycorrhizal fungi
(Shearer and Kohl 1986; Handley et al. 1993). In
seven of the 61 solvable models (11 % of total)
species 9 plot solutions included nitrate proportions
that exceeded the 100 % upper fractional limit of N
use. Because these violations were small and unidi-
rectional (average of 0.15 units above (1) we forced
these nitrate fractional contributions to ‘1’, read as
100 % of tree N use, for the sake of graphical
analyses. Omission of these solutions did not alter
statistical conclusions and both model parameters and
outputs as well as the primary data are reported in
Supplementary Tables S1–S4.
Earlier studies in Panama found very little fine root
biomass [0.2 m deep (Yavitt and Wright 2001), and
the high densities of roots located in surface soils were
unaltered by N or P (Yavitt et al. 2011). This suggests
minimal access to deeper pools of N. However,
previous observations found soil nitrate at depth to
be as enriched as our DON values (Koehler et al.
2012). As such, their incorporation into modeling
efforts would place plant d15N end members outside of
soil source mixtures (comparable to DON sources),
requiring a large (and unsupported by literature
estimates) fractionation magnitude to force plant
d15N values into solvable ranges for mixing models.
As both nitrate and ammonium concentrations also
drop precipitously with depth (Koehler et al. 2012),
the solutions presented from surface N sources are
therefore likely reasonable approximations of N
accessed by these tree species.
Biogeochemistry (2014) 119:293–306 297
123
Results
The response of soil nutrient and d15N values to N
and P fertilization
Soil nitrate concentrations from in situ ion exchange
resin bags increased across all three rainy season
deployments in response to N addition, although the
increase was only marginally significant in the initial
incubation (pMCMC = 0.09, 0.002, 0.0001, respec-
tively; Table 1). Similarly, P addition increased resin
adsorbed phosphate across all three rainy season
measurements (pMCMC = 0.055, 0.007, 0.0008,
respectively; Table 1). Only resin nitrate concentra-
tions from the late rainy season exhibited an N 9 P
interaction (pMCMC = 0.025) due to lower nitrate
concentrations in the ?N?P treatment than in control
plots. In contrast, there were no significant effects of N
or P addition on resin ammonium for the same three
rainy season deployments or on concentrations of any
N form measured in the KCl extracts.
Soil nitrate d15N values obtained from the KCl
extraction and the fully replicated mid-rainy season
resin incubation were significantly enriched following
N addition (p = 0.004, 0.025, respectively; Table 2).
There was also a marginally significant N 9 P inter-
action in the mid-rainy season resin nitrate d15N
values (pMCMC = 0.058) due to d
15N values in the
?N?P treatment being lower than the ?N treatment.
Soil d15N of DON was significantly greater than
ammonium (p \ 0.001, paired Wilcoxon signed-rank
test), but neither values were altered by fertilization
(p [ 0.05). The d15N and d18O of soil nitrate from salt
extracts were positively correlated with one another in
?N and ?N?P treatments (r2 = 0.63, p \ 0.001,
n = 17, and 0.44, p \ 0.001, n = 18, respectively)
but were not correlated in the control or ?P treatments
(p [ 0.2; Supplementary Fig. S2).
Response of foliar and litterfall d15N values to N
and P addition
Foliar d15N values responded significantly to N
fertilization in two of the tree species; the overall N
addition effect increased foliar d15N values in A.
blackiana (pMCMC = 0.017), and H. concinna became
enriched in response to both N and P additions
(pMCMC = 0.023, 0.017). Despite having no overall N
or P effect in O. mapora and T. panamensis, foliar
d15N values of these trees exhibited significant or
marginally significant N 9 P interactions owing to
distinctively higher d15N values in the ?N relative to
?N?P plots (pMCMC = 0.018, 0.069, respectively;
Table 1 Soil mineral nutrient concentrations after 13 years of fertilization
C N NP P
2 M KCl (lg g-1)
[NH4
?] 2.20 ± 0.80 1.44 ± 0.14 1.56 ± 0.65 2.47 ± 0.48
[NO3
-] 4.18 ± 0.81 4.98 ± 1.07 5.44 ± 0.85 3.32 ± 0.44
[DON] 72.1 ± 10.7 85.1 ± 9.50 70.0 ± 11.8 83.2 ± 15.1
Exchange resin (lg-1 mL-1 g-1 day)
Early [NH4
?] 1.19 ± 0.28 0.84 ± 0.16 0.43 ± 0.06 0.83 ± 0.45
Mid [NH4
?] 1.28 ± 0.26 2.53 ± 1.01 1.76 ± 0.49 1.13 ± 0.20
Late [NH4
?] 2.01 ± 0.22 2.69 ± 0.49 1.36 ± 0.25* 2.44 ± 0.31
Early [NO3
-] 4.83 ± 2.01 10.21 – 3.43 9.23 – 1.02 5.42 ± 2.33
Mid [NO3
-] 13.27 ± 5.09 27.86 – 4.12 15.96 – 2.09 6.38 ± 1.83
Late [NO3
-] 14.27 ± 5.90 42.98 – 10.63 13.39 – 2.41* 9.97 ± 1.89
Early [PO4
-] 0.02 ± 0.01 0.03 ± 0.02 1.28 – 0.18 1.47 – 0.59
Mid [PO4
-] 0.45 ± 0.23 1.37 ± 1.10 4.44 – 1.19 3.62 – 0.57
Late [PO4
-] 0.08 ± 0.03 0.02 ± 0.01 4.12 – 0.87 4.45 – 1.56
2 M KCl extractions and in situ ion exchange resin bags during early, mid, and late periods of the 2011 rainy season, the thirteenth
year of fertilization. Bold values represent a significant N or P addition effect (p B 0.10) except where there was a significant N 9 P
interaction (indicated with an * p B 0.10)
298 Biogeochemistry (2014) 119:293–306
123
Fig. 1a). These changes to foliar d15N values were not
accompanied by changes in leaf N concentrations
(Mayor et al. 2014) nor was variability within species
significantly altered by treatments (p [ 0.05, Le-
vene’s test). Of the species by treatment combinations,
the largest range of foliar d15N values was found for A.
blackiana in ?N?P and ?P plots (2.66 and 3.04 %,
respectively), and T. panamensis in the ?N plots
(2.56 %). The range of all other species by treatment
combinations was 0.59–1.91 %.
As for leaves, the overall effect of N addition
increased leaf litterfall d15N values significantly in H.
concinna (p = 0.011) and non-significantly in A.
blackiana (Fig. 1b). Leaf d15N values, hypothesized
to be susceptible to continued accumulation of 15N-
enriched N during leaf N resorption (Evans 2001),
were unrelated to calculated N resorption efficiencies
((leaf [mg g-1] - leaf litter [mg g-1]) / leaf
[mg g-1]) 9 100 (Killingbeck 1996) across treat-
ments (Mayor et al. 2014). Furthermore, leaf litter
was not uniformly 15N-enriched in A. blackiana and H.
concinna in N addition treatments, as would be
expected if resorption of 14N contributed to green leaf
enrichment. Rather, the average difference from leaf to
litter was 0.44 % in A. blackiana (n = 7) and -0.49 %
in H. concinna (n = 7), suggesting altered resorption
dynamics did not contribute to green leaf 15N-enrich-
ment. The resorption efficiency of d15N was not
significantly altered by nutrient addition in any species
(data not shown), and is not discussed further.
15N-based mixing model solutions
Of 64 potential models (four species 9 four plots per
treatment 9 four treatments), 61 IsoError solutions
were obtained due to the absence of one tree species in
three plots (Supplementary Table S2). Of those, 89 %
were solvable using the adjusted foliar (Eq. 4) and
resin extracted soil N d15N values. As seen in Fig. 2,
average mineral N contributions suggest that trees did
not use proportionally greater amounts of either
mineral N form (e.g. average NO3
- to NH4
? use ratio
&1) in control and P addition plots. However, T.
panamensis and O. mapora exhibited slight trends
toward greater proportional use of nitrate and ammo-
nium, respectively (Fig. 2). In N addition treatments,
Table 2 Soil nitrogen d15N values after 13 years of fertilization
d15N (%) C N NP P
2 M KCl NO3
- -0.5 ± 0.5 3.6 – 1.4 2.2 – 1.2 0.1 ± 1.4
2 M KCl DON 9.6 ± 0.4 9.6 ± 0.6 9.0 ± 0.6 8.9 ± 0.1
Early resin NO3
- -1.0 ± 0.7 0.1 ± 1.7 -1.1 ± 0.1 -0.1 ± 0.5
Mid resin NO3
- -0.8 ± 0.7 0.0 – 1.0 -1.4 – 0.7* -1.2 ± 0.7
Mid resin NH4
? 3.9 ± 1.2 5.6 ± 0.1 4.4 ± 0.8 3.6 ± 1.1
2 M KCl extractions and in situ ion exchange resin bags during early and mid-rainy season incubation periods of the 2011 rainy
season, the thirteenth year of fertilization. The d15N values of soil N forms were measured using the denitrifier method and were fully
replicated for both ions across all 16 plots only for the mid-season resin bags. Bold values represent a significant N or P addition
effect (p \ 0.05) except where there was a significant N 9 P interaction (indicated with an * p \ 0.10)
Fig. 1 Leaf and litter d15N values from four locally common
tree species in the 13th year of fertilization. Fertilization
treatments were nitrogen (N), phosphorus (P), or both (NP).
Tree species: ALSEBL, Alseis blackiana; HEISCO, Heisteria
concinna; OENOMA, Oenocarpus mapora; TET2PA, Tetra-
gastris panamensis. Red letters indicate significant (p B 0.05)
overall N or P fertilization effects or N 9 P interactions where
N effects were contingent upon P co-addition. (Color figure
online)
Biogeochemistry (2014) 119:293–306 299
123
however, proportional use of nitrate increased in all
species, significantly so in H. concinna (p = 0.062),
T. panamensis (p = 0.021), and O. mapora
(p = 0.015) calculated separately, and across all
species when analyzed together (p = 0.027). Nitrate
use was slightly reduced under P co-addition, but not
enough to elicit an N 9 P interaction.
Across treatments, average contributions of soil
nitrate were positively related to all three in situ
metrics of soil nitrate availability (r2 = 0.22, 0.45,
0.51; n = 16, 16, and 15, respectively; Fig. 3), but not
ammonium. More specifically, mixing model solu-
tions suggest that N addition caused three of the four
tree species to increase proportional use of soil nitrate
by threefold in ?N versus –N treatments
(f NO3=f NH4 = 3.7 vs. 1.17, respectively; Supplemen-
tary Table S3). Time-integrated metrics of bioavail-
able soil N data exhibited similar, but lower fractional
increases in [NO3
-]:[NH4
?] ratios (13.26 vs. 6.3 in
?N vs. –N treatments, respectively). Incorporation of
these relative concentrations as weightings (see
Eq. [13] in Phillips and Koch (2002)) in mass balance
models increased proportional nitrate use estimates
across all plots, thus diminishing the relative ?N
addition effect from a 40 % increase in nitrate use to
15 % (Supplementary Table S3).
Discussion
The effects of altered N cycling on soil d15N
Nitrogen addition more than doubled nitrate accumu-
lated on ion exchange resins and this nitrate pool was
15N-enriched relative to plots that did not receive N. In
contrast, N addition did not affect the concentration or
d15N values of ammonium or DON (Table 1, 2), nor
were total soil N concentrations or d15N values
modified following 8 years of N addition (Koehler
et al. 2009). The increased concentrations of extract-
able nitrate, but not ammonium, in N addition plots
were in agreement with a previous study 4 years after
the experiment began (Yavitt et al. 2011) and in a
monthly sampling effort in year ten (Turner et al.
2013). The increased concentration of nitrate, but not
ammonium, is likely caused by rapid dissolution of the
urea fertilizer and nitrification of resulting ammonium.
This is in agreement with observations in the four ?N
Fig. 2 Average 15N-based
mixing model solutions
estimating each tree species’
proportional use (as % of
total leaf N) of inorganic N
sources following 13 years
of fertilization. Error bars
are the standard error from a
mean of means from
treatment replicates. Nitrate,
black; ammonium, grey.
Arrows mean nitrate use.
(Color figure online)
300 Biogeochemistry (2014) 119:293–306
123
plots that 9 years of N addition increased mineraliza-
tion, nitrification, and denitrification rates in this
experiment (Corre et al. 2010).
Isotopic evidence from this and a previous study
(Koehler et al. 2012) suggest that increased trace gas
losses resulting from the greater soil nitrate concen-
trations contributed to the observed 15N-enrichment of
the soil nitrate pool. The remaining nitrate pool
becomes 15N- and 18O-enriched because the lighter
isotopomers are enzymatically favored by denitrifiers
(Pe´rez et al. 2006; Wunderlich et al. 2013). In the
Gigante Fertilization Experiment, positive correla-
tions between d15N and d18O values of soil nitrate
following N addition provide evidence for the imprint
of gaseous losses on soil nitrate (Supplementary Fig.
S2), and isotopomer maps found clear imprints of N2O
production from denitrification rather than
nitrification (Fig. 5 in Koehler et al. 2012). Positive
correlations between d15N and d18O values of soil
nitrate are due to fractionation of both isotopes during
denitrification (Houlton et al. 2006; Po¨rtl et al. 2007;
Park et al. 2011), a small-scale process that translates
to regional and globally enriched tropical soil d15N
values in modeling exercises (Bai and Houlton 2009;
Bai et al. 2011). That nitrous oxide emissions are
stimulated by N additions to tropical forests is in
agreement with previous studies (Hall and Matson
2003; Lohse and Matson 2005; Corre et al. 2010; Lu
et al. 2011) and confirms the theorized mechanism
accounting for regional enrichment of canopy leaf
d15N values in Panama over decadal time scales (Hietz
et al. 2011).
Urea volatilization, a potentially fractionating pro-
cess that could temporarily cause 15N-enrichment of
ammonium pools (Marshal et al. 2007; Khanif 1992),
was unlikely in our system for several reasons. First,
wet season ammonium was not isotopically enriched
in this or a previous study (Koehler et al. 2012).
Second, ammonia volatilization is typically observed
under low rainfall conditions (Bouwmeester et al.
1985) where urea is not given a chance to dissolve,
disassociate, and diffuse into the subsoil (Christianson
et al. 1993). Third, ammonia volatilization occurs
under basic soil pH conditions. None of these condi-
tions characterize this experiment; urea was dissolved
prior to wet season application, and its addition
lowered soil pH values relative from 5.25 in control
plots to 4.47 in ?N and 4.85 in ?N?P plots
(p \ 0.001; Turner et al. 2013).
The effects of altered P cycling on soil d15N values
Phosphorus addition increased resin phosphate regard-
less of N co-addition (Table 1), which agrees with
previous findings using different extraction methods
(Turner et al. 2013; Yavitt et al. 2011). Interestingly, P
addition also influenced soil nitrate d15N indirectly, as
indicated by the N 9 P interaction for the fully
replicated mid-rainy season resin nitrate (Table 2).
This was caused by the loss of 15N-enrichment in plots
receiving both N and P. The observed 15N-enrichment
under N addition, but not under N and P co-addition,
suggests P co-addition directly or indirectly reduced
fractionating losses caused by N addition, perhaps
through an increase in microbial competition for soil
nitrate prior to denitrification or a decrease in
Fig. 3 Relationship between foliar 15N-based mixing model
solutions of nitrate use and nitrate concentrations of in situ ion
exchange resin bags during the 2011 rainy season. Higher
proportional use of nitrate corresponded to higher nitrate
concentrations at all time periods (r2 = 0.22, 0.45, 0.51;
n = 16, 16, and 15, respectively; n = 16, respectively). Axes
reversed for graphical purposes. Each symbol is the average of
model solutions for four tree species (one outlier from late rainy
season omitted) across control (open circle), ?P (open square),
?N (filled circle), and ?N?P (filled square) treatments
Biogeochemistry (2014) 119:293–306 301
123
nitrification limiting nitrate supply to denitrifiers.
Support for the first possibility includes the observa-
tion that microbial biomass was increased by P
addition after 10 years of treatments (Turner et al.
2013). Support for the second possibility are observa-
tions following 11 years of N-addition to P-limited
Hawaiian rainforest, where overall N-oxide emissions
were more strongly related to rates of nitrification than
of denitrification (Hall and Matson 2003).
The influence of altered soil nutrient availability
on leaf d15N values
The variation in foliar d15N values in control plots
(maximum range = 4.37 %, 90 % quantile
range = 3.27 %) was similar to or smaller than more
exhaustive taxonomic sampling from other lowland
tropical forests (Koba et al. 2010; Guehl et al. 1998;
Martinelli et al. 1999), but equivalent to the 90 %
quantile range of exhaustive (n = 213 species) sam-
pling of ‘‘sun-leaves’’ on nearby Barro Colorado
Island (maximum range = 9.38 %, whereas 90 %
quantile range = 3.31 %; subset of data from Hietz
et al. 2011).
Of the four locally common tree species, A.
blackiana and H. concinna produced 15N-enriched
leaf material in N addition plots (Fig. 1a), a pattern
that parsimoniously suggests access to the enriched
nitrate pool in the same treatments. Foliar N concen-
trations, as detailed in a companion manuscript, were
unresponsive to N addition in all four of these tree
species (Mayor et al. 2014). However, P addition also
influenced foliar d15N values: H. concinna leaves were
enriched in response to P addition, and the trend
toward 15N-enrichment in O. mapora and T. panam-
ensis leaves was suppressed in ?N?P plots (Fig. 1a).
These N 9 P interactive patterns were similar to the
suppression of nitrate 15N-enrichment in ?N?P plots
(Table 2). Collectively, these patterns suggest that the
use of foliar d15N to indicate the openness of a tropical
forest N cycle depends in part on how soil N cycling
and plant mineral N use respond to P availability. In
particular, d15N values from tropical trees growing on
soils relatively high in P may be less isotopically
sensitive to N deposition than trees growing on
relatively P poor soils. Similarly, 2 years of P addition
to an arctic ecosystem increased foliar d15N values
significantly in two common plant species, but not in
two rare species; a pattern attributed to increased N
mineralization following P addition (Giblin et al.
1991, Nadelhoffer et al. 1996).
Similar to canopy foliage, litterfall d15N values
increased in response to both N and P additions in A.
blackiana and H. concinna, although only signifi-
cantly so for H. concinna in N addition plots, in part
due to composited sampling reducing within-plot
statistical power. In a previous litterfall study, com-
munity-wide fine litterfall was approximately 1 %
enriched after ten years of N addition (Corre et al.
2010). These results demonstrate: (i) a clear pathway
by which the isotopic imprint of some tree species can
feedback to soils as annual N application rates were
comparable to annual litterfall inputs (Yavitt et al.
2011); (ii) the importance of sampling litterfall from
more than one species unless it is predetermined to be
representative of the wider community (i.e. H.
concinna); and, (iii) that litterfall may be a suitable
substitute for relative comparisons when sampling
canopy foliage is impossible to obtain.
The d15N based mass balance models indicated that
the proportional use of nitrate increased in accordance
with increased nitrate concentrations following
chronic N addition (Fig. 3; Supplementary Table
S2). This interpretation was supported by simple
observations of enrichment patterns as well as statis-
tical probabilities derived from mass balance mixing
models based on d15N values alone and with respec-
tive mineral N concentrations as weightings. Including
nutrient weightings, however, diminished the magni-
tude of change from a 40 % relative increase in nitrate
use following N addition to a 15 % increase (Supple-
mentary Table S3). While use of concentration
weightings from time-integrated resins has certain
intuitive appeal, their acceptance requires two main
assumptions. First, the resin nitrate to ammonium
ratios must be seen as a precise metric of their
respective bioavailability experienced by trees. Sec-
ond, any a priori bias toward one mineral N form, such
as through transporter affinities or ion-influx interac-
tions between [NO3
-] and [NH4
?] during N absorp-
tion (Glass et al. 2002; Bassirirad 2000), must be
accounted for. Regardless, both approaches indicated
that tropical trees increased proportional use of nitrate
in response to its greater relative availability resulting
from chronic N addition. As a result, there was no
strong evidence that preference for ammonium, for
instance, was retained irrespective of its relative
302 Biogeochemistry (2014) 119:293–306
123
availability. This finding is in agreement with previous
observations that plants track nitrate availability in
lower latitude ecosystems (Houlton et al. 2007; Wang
and Macko 2011; Kahmen et al. 2008; Takebayashi
et al. 2010).
Although tree use of DON has been shown using
d15N data from montane and high-latitude ecosystems
(Averill and Finzi 2011; Mayor et al. 2012), there was
no isotopic evidence suggesting the tropical trees used
DON in our system or along a precipitation gradient in
Hawaiian tropical montane forest (Houlton et al.
2007). Although use of 15N-labeled amino acids has
been demonstrated in tropical montane palms (Ander-
sen and Turner 2013), appreciable use of such
compounds is likely not beneficial to plants growing
on soils with highly mobile mineral nitrogen concen-
trations. Indeed, although data from tropical forests
are exceedingly rare, studies from a variety of
montane tropical forest plants suggests that mineral
N forms are often preferred over amino acids even
when provided in concert at comparable concentra-
tions (Andersen K, Mayor J, Turner B, unpublished
data). Furthermore, the predominance of arbuscular
mycorrhizal associations among most tropical tree
species (including those in this study) precludes the
enzymatic advantage described for ectomycorrhizal
associations that retain the capacity for the degrada-
tion of various organic N compounds in N-limited
ecosystems (Talbot and Treseder 2010).
Conclusion
Tropical regions account for 71 % of global N losses
from unmanaged land and the continued gaseous loss
of isotopically light nitrous oxides leaves an isotopic
imprint on the residual nitrate pool from which it is
derived (Bai et al. 2011). Therefore, the mechanism
used to account for the gradual increase in archived
plant d15N samples seen in tropical forests at decadal
timescales (Hietz et al. 2011) is now empirically
supported. This study, from the longest running
fertilization experiment in lowland tropical forest,
demonstrates that despite chronic addition of isotopi-
cally depleted urea, rapid rates of N cycling resulted in
relatively higher concentrations of nitrate that com-
prised greater proportions of tree N use. Surprisingly,
the relative availability of P also altered the isotopic
imprint in both soils and plants, possibly through
increased microbial demand for N under P addition.
Coexistence of the tree species examined here appear
to not be linked to specialization on N forms; instead,
the sampled plant community appears to prefer
roughly equal proportions of ammonium and nitrate
under ambient conditions, and continued anthropo-
genic inputs of N will likely result in accelerated N
cycling, and increased use of nitrate by all species.
Acknowledgments Support for J.M. was provided by the
National Science Foundation’s International Research
Fellowship Program (OISE-1012703) and in-kind support from
the Smithsonian Tropical Research Institute’s Fellowship
Program. We thank Julio Rodriquez, Dayana Agudo, Helene
Mueller-Landau, Luis Ramos, Rueben Hernandez, Sarah Dale,
and Tania Romero for their contributions, and the Republic of
Panama for providing access to their natural resources for
scientific research. We thank Grace Crummer, Jason Curtis, and
Kathy Curtis for their contributions at the University of Florida.
JM, BT, and JW designed the experiment and wrote the
manuscript; TS contributed analytical ideas and logistical
support, MB contributed a statistical method. The manuscript
was improved due to the untiring efforts of three anonymous
reviewers.
References
Amundson R, Austin AT, Schuur EAG, Yoo K, Matzek V,
Kendall C, Uebersax A, Brenner D, Baisden WT (2003)
Global patterns of the isotopic composition of soil and
plant nitrogen. Glob Biogeochem Cycle 17:31
Andersen KA, Turner BL (2013) Preferences or plasticity in
nitrogen acquisition by understory palms in a tropical
montane forest? J Ecol 101:819–825
Ashton IW, Miller AE, Bowman WD, Suding KN (2008)
Nitrogen preferences and plant–soil feedbacks as influ-
enced by neighbors in the alpine tundra. Oecologia
156:625–636
Averil C, Finzi A (2011) Increasing plant use of organic nitrogen
with elevation is reflected in nitrogen uptake rates and
ecosystem dN15. Ecology 92:883–891
Baayan RH, Davidson DJ, Bates DM (2008) Mixed-effects
modeling with crossed random effects for subjects and
items. J Mem Lang 59:390–412
Baayen RH (2011) languageR: Data sets and functions with
‘‘Analyzing Linguistic Data: A practical introduction to
statistics’’. R package version 14 http://CRAN.R-project.
org/package=languageR
Bai E, Houlton BZ (2009) Coupled isotopic and process-based
modeling of gaseous nitrogen losses from tropical rain
forest. Global Biogeochem Cycle 23 (GB2011)
Bai E, Houlton BZ, Wang Y (2011) Isotopic identification of
global nitrogen hotspots across natural terrestrial ecosys-
tems. Biogeosci Discuss 8:12113–12152
Bassirirad H (2000) Kinetics of nutrient uptake by roots:
responses to global change. New Phytol 147:155–169
Biogeochemistry (2014) 119:293–306 303
123
Bates D, Maechler M, Bolker B (2012) lme4: Linear mixed-
effects models using S4 classes. R package version
0.999999-0. http://CRANR-project.org/package=lme4
Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR,
Stevens HH, White J-SS (2009) Generalized linear mixed
models: a practical guide for ecology and evolution. Trends
Ecol Evol 24:127–135
Bouwmeester RJB, Vlek PLG, Stumpe JM (1985) Effect of
environmental factors on ammonia volatilization from a
urea-fertilized soil. Soil Sci Soc Am J 49:376–381
Brookshire ENJ, Hedin LO, Newbold JD, Sigman DM, Jackson
JK (2012) Sustained losses of bioavailable nitrogen from
montane tropical forests. Nat Geosci 5:1–4
Cabrera ML, Beare MH (1993) Alkaline persulfate oxidation for
determining total nitrogen in microbial biomass extracts.
Soil Sci Soc Am J 57:1007–1012
Christianson CB, Baethgen WE, Carmona G, Howard RG
(1993) Microsite reactions of urea-nbtpt fertilizer on the
soil surface. Soil Biol Biochem 25:1107–1117
Corre MD, Veldkamp E, Arnold J, Wright SJ (2010) Impact of
elevated N input on soil N cycling and losses in old-growth
lowland and montane forests in Panama. Ecology
91:1715–1729
Davidson EA, Reis de Carvalho CJ, Figueira AM, Ishida FY,
Ometto JPHB, Nardoto GB, Saba RT, Hayashi SN, Leal
EC, Vieira ICG, Martinelli LA (2007) Recuperation of
nitrogen cycling in Amazonian forests following agricul-
tural abandonment. Nature 447:995–998
Dawson TE, Mambelli S, Plamboeck AH, Templer PH, Tu KP
(2002) Stable isotopes in plant ecology. Annu Rev Ecol
Syst 33:507–559
Dijkstra P, LaViolette CM, Coyle JS, Doucett RR, Schwartz E,
Hart SC, Hungate BA (2008) 15N enrichment as an inte-
grator of the effects of C and N on microbial metabolism
and ecosystem function. Ecol Lett 11:1–9
Doyle A, Weintraub MN, Schimel JP (2004) Persulfate diges-
tion and simultaneous colorimetric analysis of carbon and
nitrogen in soil extracts. Soil Sci Soc Am J 68:669–676
Evans RD (2001) Physiological mechanisms influencing plant
nitrogen isotope composition. Trends Plant Sci 6:121–126
Giblin AE, Nadelhoffer KJ, Shaver GR, Laundre J, McKerrow
A (1991) Biogeochemical diversity along a topographic
gradient in a tundra landscape. Ecol Monogr 61:415–436
Giblin AE, Laundre JA, Nadelhoffer KJ, Shaver GR (1994)
Measuring nutrient availability in arctic soils using ion
exchange resins: a field test. Soil Sci Soc Am J
58:1154–1162
Glass ADM, Britto DT, Kaiser BN, Kinghorn JR, Kronzucker HJ,
Kumar A, Okamoto M, Rawat S, Siddiqi MY, Unkles SE,
Vidmar JJ (2002) The regulation of nitrate and ammonium
transport systems in plants. J Exp Bot 53:855–864
Guehl JM, Domenach AM, Bereau M, Barigah TS, Casabianca
H, Ferhi A, Garbaye J (1998) Functional diversity in an
Amazonian rainforest of French Guyana: a dual isotope
approach (d15N and d13C). Oecologia 116:316–330
Hall SJ, Matson PA (2003) Nutrient status of tropical rain forests
influences soil N dynamics after N additions. Ecol Monogr
73:107–129
Handley LL, Daft MJ, Wilson J, Scrimgeour CM, Ingleby K,
Sattar MA (1993) Effects of the ecto-mycorrhizal and VA-
mycorrhizal fungi Hydnagium carneum and Glomus
clarum on the Delta-N-15 and Delta-C-13 values of
Eucalyptus globulus and Ricinus communis. Plant Cell
Environ 16:375–382
Harrison KA, Bol R, Bardgett RD (2007) Preferences for dif-
ferent nitrogen forms by coexisting plant species and soil
microbes. Ecology 88:989–999
Hedin LO, Brookshire ENJ, Menge DNL, Barron AR (2009)
The nitrogen paradox in tropical forest ecosystems. Annu
Rev Ecol Syst 40:613–635
Hietz P, Du¨nisch O, Wanek W (2010) Long-term trends in
nitrogen isotope composition and nitrogen concentration in
Brazilian rainforest trees suggest changes in nitrogen cycle.
Environ Sci Technol 44:1191–1196
Hietz P, Turner BL, Wanek W, Richter A, Nock CA, Wright SJ
(2011) Long-term change in the nitrogen cycle of tropical
forests. Science 334:664–666
Hobbie EA, Ho¨gberg P (2012) Nitrogen isotopes link mycor-
rhizal fungi and plants to nitrogen dynamics. New Phyt
196:367–382
Hobbie EA, Ouimette AP (2009) Controls of nitrogen isotope
patterns in soil profiles. Biogeochemistry 95:355–371
Hobbie EA, Jumpponen A, Trappe J (2005) Foliar and fungal
15N:14N ratios reflect development of mycorrhizae and
nitrogen supply during primary succession: testing ana-
lytical models. Oecologia 146:258–268
Houlton BZ, Sigman DM, Hedin LO (2006) Isotopic evidence
for large gaseous nitrogen losses from tropical rainforests.
Proc Nat Acad Sci USA 103(23):8745–8750
Houlton BZ, Sigman DM, Schuur EAG, Hedin LO (2007) A
climate-driven switch in plant nitrogen acquisition within
tropical forest communities. Proc Nat Acad Sci USA
104:8902–8906
IAEA (2004) IAEA analytical quality control services reference
materials catalouge 2004–2005. IAEA, Vienna
Kahmen A, Wanek W, Buchmann N (2008) Foliar d15N values
characterize soil N cycling and reflect nitrate or ammonium
preference of plants along a temperate grassland gradient.
Oecologia 156:861–870
Khanif YM (1992) Ammonia volatilization from Malaysian
soils following application of urea. Pertanika 15:115–120
Killingbeck KT (1996) Nutrients in senesced leaves: keys to the
search for potential resorption and resorption proficiency.
Ecology 77:1716–1727
Knapp AN, Sigman DM, Lipschultz F (2005) N isotopic com-
position of dissolved organic nitrogen and nitrate at the
Bermuda Atlantic Time-series Study site. Glob Biogeo-
chem Cycle 19:GB1018
Koba K, Isobe K, Takebayashi Y, Fang YT, Sasaki Y, Saito W,
Yoh M, Mo J, Liu L, Lu X, Zhang T, Zhang W, Senoo K
(2010) d15N of soil N and plants in a N-saturated, sub-
tropical forest of southern China. Rapid Commun Mass
Spectrom 24:2499–2506
Koehler B, Corre MD, Veldkamp E, Wullaert H, Wright SJ
(2009) Immediate and long-term nitrogen oxide emissions
from tropical forest soils exposed to elevated nitrogen
input. Glob Change Biol 15:2049–2066
Koehler B, Corre MD, Steger K, Well R, Zehe E, Sueta JP,
Veldkamp E (2012) An in-depth look into a tropical low-
land forest soil: nitrogen-addition effects on the contents of
N2O, CO2 and CH4 and N2O isotopic signatures down to
2-m depth. Biogeochemistry 111:695–713
304 Biogeochemistry (2014) 119:293–306
123
Kramer MG, Sollins P, Sletten RS, Swart PK (2003) N isotope
fractionation and measures of organic matter alteration
during decomposition. Ecology 84:2021–2025
Lohse KA, Matson P (2005) Consequences of nitrogen additions
for soil losses from wet tropical forests. Ecol Appl
15:1629–1648
Lu M, Yang Y, Luo Y, Fang C, Zhou X, Chen J, Yang X, Li B
(2011) Responses of ecosystem nitrogen cycle to nitrogen
addition: a meta-analysis. New Phyt 189:1040–1050
Marshal JD, Brooks JR, Lajtha K (2007) Sources of variation in
the stable isotopic composition of plants. In: Michener R
(ed) Stable isotopes in ecology and environmental science.
Blackwell Publishing, USA, pp 22–60
Martinelli LA, Piccolo MC, Townsend AR, Vitousek PM,
Cuevas E, McDowell W, Robertson GP, Santos OC, Tre-
seder K (1999) Nitrogen stable isotopic composition of
leaves and soil: tropical versus temperate forests. Biogeo-
chemistry 46:45–65
Mayor JR, Schuur EAG, Mack MC, Hollingsworth TN, Ba˚a˚th E
(2012) Nitrogen isotope patterns in Alaskan black spruce
reflect organic nitrogen sources and the activity of ecto-
mycorrhizal fungi. Ecosystems 15:819–831
Mayor JR, Wright SJ, Turner BL (2014) Species-specific
responses of foliar nutrients to long-term nitrogen and
phosphorus additions in a lowland tropical forest. J Ecol-
ogy 102:36–44
Nadelhoffer K, Shaver G, Fry B, Giblin A, Johnson L, McKane
R (1996) 15N natural abundances and N use by tundra
plants. Oecologia 107:386–394
Niklas KJ, Owens T, Reich PB, Cobb ED (2005) Nitrogen/
phosphorus leaf stoichiometry and the scaling of plant
growth. Ecol Lett 8:636–642
Pardo LH, Nadelhoffer KJ (2010) Using nitrogen isotope
ratios to assess terrestrial ecosystems at regional and
global scales. In: West J, Bowen GJ, Dawson TE, Tu KP
(eds) Isoscapes: understanding movement, pattern, and
process on Earth through isotope mapping. Springer,
New York
Park S, Pe´rez T, Boering KA, Trumbore SE, Gil J, Marquina S,
Tyler SC (2011) Can N2O stable isotopes and isotopomers
be useful tools to characterize sources and microbial
pathways of N2O production and consumption in tropical
soils? Glob Biogeochem Cycle 25:GB1001
Pe´rez T, Trumbore SE, Tyler SC, Davidson EA, Keller M, de
Camargo PB (2000) Isotopic variability of N2O emissions
from tropical forest soils. Glob Biogeochem Cycle
14:525–535
Pe´rez T, Garcia-Montiel D, Trumbore S, Tyler S, Camargo P,
Moreira M, Piccolo M, Cerri C (2006) Nitrous oxide
nitrification and denitrification 15N enrichment factors
from amazon forest soils. Ecol Appl 16:2153–2167
Phillips DL, Koch PL (2002) Incorporating concentration
dependence in stable isotope mixing models. Oecologia
130:114–125
Phillips DL, Newsome SD, Gregg JW (2005) Combining
sources in stable isotope mixing models: alternative
methods. Oecologia 144:520–527
Po¨rtl K, Zechmeister-Boltenstern S, Wanek W, Ambus P, Ber-
ger TW (2007) Natural 15N abundance of soil N pools and
N2O reflect the nitrogen dynamics of forest soils. Plant Soil
295:79–94
Santiago LS, Wright SJ, Harms KE, Yavitt JB, Korine C, Garica
MN, Turner BL (2012) Tropical tree seedling growth
responses to nitrogen, phosphorus and potassium addition.
J Ecol 100:309–316
Schimann H, Pontona S, Hattenschwilerb S, Ferryc B, Lensib R,
Domenacha A-M, Roggy J-C (2008) Differing nitrogen use
strategies of two tropical rainforest late successional tree
species in French Guiana: evidence from 15N natural
abundance and microbial activities. Soil Biol Biochem
40:487–494
Schuur EAG, Matson PA (2001) Aboveground net primary
productivity and nutrient cycling across a mesic to wet
precipitation gradient in Hawaiian montane forest. Oeco-
logia 128:431–442
Shearer G, Kohl DH (1986) N2 fixation in field settings, esti-
mations based on natural 15N abundance. Aust J Plant Phys
13:699–757
Sigman DM, Casciotti KL, Andreani M, Barford C, Galanter M,
Bohlke JK (2001) A bacterial method for the nitrogen
isotopic analysis of nitrate in seawater and freshwater. Anal
Chem 73:4145–4153
Takebayashi Y, Koba K, Sasaki Y, Fang Y, Yoh M (2010) The
natural abundance of 15N in plant and soil-available N
indicates a shift of main plant N resources to NO3
- from
NH4
? along the N leaching gradient. Rapid Comm Mass
Spectrom 24:1001–1008
Talbot JM, Treseder KK (2010) Controls over mycorrhizal
uptake of organic nitrogen. Pedobiologia 53:169–179
Templer PH, Weathers KC (2011) Use of mixed ion exchange
resin and the denitrifier method to determine isotopic val-
ues of nitrate in atmospheric deposition and canopy
throughfall. Atmos Environ 45:2017–2020
Turner BL, Wright SJ (2014) The response of microbial biomass
and hydrolytic enzymes to a decade of nitrogen, phos-
phorus, and potassium addition in a lowland tropical rain
forest. Biogeochemistry 117:115–130
Turner BL, Yavitt JB, Harms KE, Garcia MN, Romero TE,
Wright IJ (2013) Seasonal changes and treatment effects on
soil inorganic nutrients following a decade of fertilizer
addition in a lowland tropical forest. Soil Sci Soc Am J
17:1357–1369
Wang L, Macko SA (2011) Constrained preferences in nitrogen
uptake across plant species and environments. Plant Cell
Environ 34:525–534
Wang L, Shaner P-JL, Macko S (2007) Foliar d15N patterns
along successional gradients at plant community and spe-
cies levels. Geophys Res Lett 34:L16403
Wright SJ, Yavitt JB, Wurzburger N, Turner BL, Tanner EVJ,
Sayer EJ, Santiago LS, Kaspari M, Hedin LO, Harms KE,
Garcia MN, Corre MD (2011) Potassium, phosphorus, or
nitrogen limit root allocation, tree growth, or litter pro-
duction in a lowland tropical forest. Ecology
92:1616–1625
Wunderlich A, Meckenstock RU, Einsiedl F (2013) A mixture
of nitrite-oxidizing and denitrifying microorganisms
affects the d18O of dissolved nitrate during anaerobic
microbial denitrification depending on the d18O of ambi-
ent water. Geochim Cosmochim Ac 119:31–45
Yavitt JB, Wright SJ (2001) Drought and irrigation effects on
fine root dynamics in a tropical moist forest, Panama.
Biotropica 33:421–434
Biogeochemistry (2014) 119:293–306 305
123
Yavitt JB, Harms KE, Garcia MN, Wright SJ, He F, Mirabello
MJ (2009) Spatial heterogeneity of soil chemical properties
in a lowland tropical moist forest, Panama. Aust J Soil Res
47:674–687
Yavitt JB, Harms KE, Garcia MN, Marabello MJ, Wright SJ
(2011) Soil fertility and fine root dynamics in response to
4 years of nutrient (N, P, K) fertilization in a lowland
tropical moist forest, Panama. Aust Ecol 36:433–445
306 Biogeochemistry (2014) 119:293–306
123