Diplomarbeit Mechanisms of nutrient dynamics in the canopy of a tropical lowland rain forest zur Erlangung des akademischen Grades Magistra der Naturwissenschaften an der Fakultät für Lebenswissenschaften der Universität Wien eingereicht von Nina Hinko Wien, Oktober 2007
58
Embed
Diplomarbeit - · PDF fileland use change and change in atmospheric composition and deposition can be critical for ... internal fluxes of nutrients within ... that of N limits
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Diplomarbeit
Mechanisms of nutrient dynamics in the canopy
of a tropical lowland rain forest
zur Erlangung des akademischen Grades
Magistra der Naturwissenschaften
an der
Fakultät für Lebenswissenschaften
der Universität Wien
eingereicht von
Nina Hinko
Wien, Oktober 2007
Index
1. General Introduction 1
Tropical rain forests 1
Nutrient cycling in tropical rain forests 1
Throughfall 4
Dry deposition 7
Canopy exchange processes 9
Study aims 14
References 15
2. Manuscript 20
Mechanisms of nutrient dynamics in the canopy of a tropical lowland rain forest
Abstract 20
Introduction 21
Materials and Methods 24
Results 31
Discussion 34
References 41
Figures and tables 45
3. Zusammenfassung 52
4. Appendix 55
- 1 -
1. General Introduction
1.1. Tropical rain forests
Tropical rain forests are distributed between the tropics of Cancer (23° 26′ 22″ N latitude) and
Capricorn (23° 26′ 22″ S latitude) and occur in all three possible tropical land areas: Africa,
South-East Asia and Australia (paleotropics) and America (neotropics) where the Americas
possess the most extensive rain forests. They exhibit an exorbitant extreme species richness
and it is estimated that approximately two thirds (170.000) of the world’s plant species occur
in tropical forests (Whitmore 1998). Further tropical rain forests are characterized by high solar
radiation because of the low latitude, a relatively stable climate throughout the year with
continuously high temperatures and high relative humidity with dry seasons being short or
negligible. Tropical rain forests therefore exhibit the highest annual rates of net primary
production (NPP) of all terrestrial ecosystems (Montagnini 2005; Jordan 1971).
Nonetheless tropical ecosystems are highly complex systems and considered to be fragile or
ecologically sensitive. Therefore effects of global change, implying not only climate change but
land use change and change in atmospheric composition and deposition can be critical for
rainforest functioning. Despite tropical rainforests exhibiting the highest biodiversity and
biomass production, most tropical soils are supposed to be highly leached and weathered and
therefore to have an extremely low fertility. Hence the major part of the nutrient pool is stored
in above-ground biomass and nutrient cycling and conservation mechanisms play a crucial role
in maintaining ecosystem productivity in tropical rain forests (Whitmore 1998).
1.2. Nutrient cycling in tropical rain forests
In addition to external nutrient inputs to and outputs from ecosystems nutrient cycling i.e.
internal fluxes of nutrients within ecosystems are thought to be crucial for sustained high
primary production of tropical rain forests. Nutrient inputs are supplied through chemical
weathering of parent material, atmospheric deposition and biological fixation of atmospheric
nitrogen. Nutrient losses occur through erosion and leaching of soils into stream- and
groundwater, emission of gases, wind-driven relocation and disturbances by natural events
such as fire. Nutrient cycling within an ecosystem is comprised of uptake of nutrients by
vegetation and microorganisms, following incorporation into organic material and finally the
release of nutrients via litterfall and throughfall and their decomposition by microorganisms. In
natural ecosystems, for instance tropical rain forests, the amount of nutrients which cycle
internally exceeds those of inputs and outputs by far and tropical rain forests therefore can be
considered as rather closed systems. In consideration of human activities, nutrient inputs and
- 2 -
outputs in ecosystems obtain an increasing significance because of additional entry of nutrients
due to fertilization, fossil fuel combustion and removal through harvest and fire (Chapin III
2002).
The productivity in terrestrial ecosystems is constrained by climate (rainfall, temperature,
seasonality, etc) and nutrient supplies mostly from the soil. Nutrients are made available for
uptake by plants through the processes of decomposition and mineralization of dead organic
matter which in turn comes from plants via litterfall, animals and microbes. Dissolved mineral
nutrients are absorbed by plants via roots and move upward in the xylem with the transpiration
water stream into leaves. There they are allocated to production of new tissue retranslocation,
or storage. Finally nutrients return to the soil via two main pathways: litterfall and throughfall
i.e. precipitation which is penetrating the canopy before it is reaching the ground. Additional
plants can trap dissolved organic nutrients e.g. nitrogen from decomposing material as soon as
they are released as well through direct transfer mediated mainly by mycorrhiza (Herrera
1978).
As mentioned above nutrient cycling is essential for continued plant growth in tropical forest
ecosystems and nutrient cycles therefore are generally found to be rather closed. Nevertheless
it must be pointed out that not all tropical forests grow on low fertility soils despite a great
proportion are oligotrophic ecosystems such as the Amazon Basin. In contrast to these nutrient
poor rain forests which receive nutrients solely from litter- and rainfall there also existing
nutrient rich tropical rain forest ecosystems especially in Central America where many soils
derived from volcanic rock. They consist of a substantial fraction of (mineral) nutrients in the
rooting zone and show nutrient cycles that are more open.
Factors that influence nutrient cycling in tropical rain forests are of biological (plant, bacterial
and fungal species), environmental (moisture and temperature) and chemical nature (nature of
soil nutrients, secondary plant components) (Attiwill 1993). Decomposition in tropical forests is
generally enhanced at high temperature, high relative moisture and availability of sufficient
oxygen which accelerate microbial activity. Hence tropical lowland rainforests exhibit a small
litter pool (Chapin III 2002; Whitmore 1998).
Litter consists of aboveground leaf litter and of belowground root litter. Soils in tropical rain
forests exhibit a remarkable amount of fine root biomass in the surface soil and litter layer
(Klinge 1973). Stark (1978) demonstrated with tracer-experiments that the fine root mat can
directly absorb dissolved nutrients released from decomposing litter and reaching the surface
soil via precipitation and throughfall. This nutrient conserving mechanism of the fine root mat
plus associated mycorrhizas enhances the uptake of nutrients, especially phosphorus (Herrera
- 3 -
1978). Nearly all tree species worldwide have ecto- or arbuscular mycorrhiza. Consequently,
the root mat plays an important role in the retention and recycling of nutrients on the forest
floor.
In tropical lowland rainforests internal recycling sustains the availability of N and P for plants
though phosphorus return via litterfall is very small compared to nitrogen (Vitousek 1984).
According to Vitousek (1982) it is assumed that P is cycled very efficient and N seems to have
low within-stand efficiency despite its rapid circulation. This supports that P availability and not
that of N limits productivity of most tropical lowland forests given that the most growth limiting
nutrient defines the cycling rate of all other nutrients (Chapin III 2002). It is not proven that N
is not used efficiently due to the fact that in those forests cycling occurs really fast and a unit of
N could circulate several times per year (Vitousek 1984). However, plant available phosphorous
is generally present at very small concentrations and is retained within the plants. Intra-plant
nutrient conservation (Jordan 1980) occurs via reabsorption and translocation of mobile
nutrients from senescing tissue through the phloem to other plant parts, i.e. to young leaves
where mineral nutrients are required to build up organic matter or to storage organs.
Nevertheless some nutrients such as Calcium are immobile in the phloem and therefore cannot
be resorbed.
Transfers of nutrients to the forest floor occur through litterfall (LF) and through precipitation in
the form of throughfall (TF) and stemflow (SF). The contribution of TF and SF as pathways for
solute (nutrient) inputs to the forest floor is important particularly in forests with abundant
rainfall and infertile soils, although amounts of nutrients in precipitation are generallly lower
than those recycled by litterfall (Tobón 2004). Nevertheless throughfall fluxes can be
ecologically significant for certain elements. In the case of potassium McDowell (1998)
demonstrated that K+ flux in TF is twice as much as in LF for a tropical wet rain forest in Puerto
Rico. In contrast nitrogen as well as calcium and phosphorus inputs to the soil derive largely
from litterfall. Calcium content in litterfall usually exceeds that of living foliage because Ca is
immobilized in cell wall components such as pectates (Johnson 1992) and can not be removed
from senescent leaves before shedding. Cavelier (1997) compared annual rates of N, P and K in
throughfall with rates in litterfall from Veneklaas (1991) for a tropical montane cloud forest. N
inputs in net throughfall NTF were significantly lower than mean inputs in litterfall while P
inputs were similar in both LF and NTF. In contrast K inputs were significantly higher in NTF
than in LF. The considerable increases of K in TF can be attributed to its high mobility and
abundance in leaves (Tobón 2004).
- 4 -
1.3. Throughfall
It is has been often demonstrated that precipitation is altered in its chemical composition
during interception through the forest canopy and that throughfall is predominantly enriched in
certain elements (Parker 1983; Lovett 1996; Veneklaas 1990; Whitmore 1998). Precipitation
which penetrates the forest canopy can reach the forest floor as throughfall (rainfall which
passed through the canopy) and as stemflow (rainfall which runs off the tree stem). Jordan
(1980) found that throughfall made up 80-90% of incident precipitation in a tropical rainforest
whereas stemflow accounted for only 1-2% of total solute fluxes. This agrees with results from
Chuyong (2004) who reported that throughfall and stemflow in a central African rain forest
accounted for 92.4-96.6% and 1.5-2.2%, respectively. Despite the very small contribution to
total nutrient fluxes stemflow is generally more enriched in nutrients than throughfall and
therefore displays a substantial yet localized nutrient input to forest floor (Parker 1983). The
difference (1.9-5.4%) between incident precipitation and TF (and SF) represents the
interception loss within the canopy. Throughfall, stemflow and interception measurements for a
tropical montane rain forest in Costa Rica resulted in 70, 2 and 28% of incident rainfall,
respectively (Hölscher 2004).
In consideration of ecosystem nutrient fluxes precipitation is a supplementary and small but
important nutrient input to forests by which nutrients from outside enter the ecosystem. For
certain elements throughfall is the major pathway for recycling and therefore the internal
nutrient dynamics of the forest (Parker 1983), particularly in wet tropical rainforests where
rainfall is high and nutrient cycling is known to be of considerable importance to meet the
nutrient demand by vegetation. Furthermore dissolved nutrients in throughfall are highly
available to organisms compared to those of litterfall, may be an important pulse at the
beginning of the rainy season and probably facilitate litter decomposition and mineralization
(Chuyong 2004) thereby reducing litter turnover times. Processes involved in chemical
alteration of precipitation are exchange processes between the canopy and the water on plant
surfaces such as leaching and uptake of dissolved material by canopy components, as well as
dissolution and wash off of dry deposited atmospheric materials such as aerosols, particulate
and gaseous substances (Lovett 1984; Hansen 1994; Parker 1983).
Parker (1983) reviewed that for K+, Na+, Mg2+ and S, the rates of NTF exceed those of litterfall
while N, P and Ca derive largely from litterfall. It is assumed that potassium derives mainly
from foliar or canopy leaching because of it is high mobility. Calcium and magnesium are also
found in higher concentrations in TF compared to BP. Calcium was shown to originate almost
equally from leaching of exchangeable pools in leaf apoplast and from wash off of particulate
- 5 -
dry deposition (Forti 1992). In contrast Mg2+, Na+ and Cl- are introduced mainly by sea salt
particles. H+ concentration (acidic deposition) in rainwater decreases during passage through
the canopy which implies that uptake (ion-exchange) processes occur. The absorption from H+
triggers increased leaching of other soluble or ion exchangeable cations such as K+, Ca2+ and
Mg2+, and weak organic acids from foliage. Numerous studies demonstrated a decline of
inorganic nitrogen concentration in precipitation during canopy penetration, especially for NH4+,
suggesting retention of N in the canopy. Apart from throughfall enrichment of mineral nutrients
it is well known that TF is additional loaded with organic compounds such as DOC and DON.
The origin of these sources is found within the canopy through leaching from canopy litter,
herbivore activities and leaching from bryophytes and exchangeable foliar pools (Coxson
1992).
Throughfall is defined as:
TF = BP + DD + CE
where TF is throughfall, BP is incident (bulk) precipitation, DD is dry deposition and CE
describes canopy exchange processes with uptake and leaching of nutrients within the canopy.
The net throughfall flux can then be calculated as:
NTF = TF – BP = DD + CE
referring to the sum of dry deposition and canopy exchange representing net throughfall
(Lovett 1996). Positive rates of NTF indicate a net enrichment of elements in throughfall due to
dry deposition and canopy leaching whereas negative rates present a net decrease in nutrient
concentration in throughfall due to canopy uptake or interception loss.
The separation between these processes, dry deposition and canopy exchange, and
quantification of their contribution to throughfall still remains a challenge. Moreover, it has
been suggested that both processes are highly dynamic (Hansen 1994) regarding to time,
space and the influence by several factors. Present data are often difficult to compares because
sampling design and analytical methods are highly variable. Collection and measurement of
bulk precipitation includes among wet deposition (elements dissolved in precipitation) also a
fraction of dry deposition such as dust and particles which enter the sampling gauge. Moreover,
dry deposition occurs to canopy surfaces causing an overestimation of canopy leaching rates
(Parker 1983). An essential part of dry deposition – deposition of gaseous substances – is
therefore not accounted for and moreover is complicated to measure. Though studies exist for
temperate forests in tropical rain forests the knowledge is even scare. Temperate forests
exhibit a less complex forest and canopy structure and are often even-sized and even-aged
compared to tropical rain forests and therefore allow an easier access for investigations.
- 6 -
The flux and chemistry of throughfall are affected by various factors and processes of
atmospheric, hydrological, chemical and biological nature, and as well human activities, which
themselves are highly temporally and spatially variable (Parker 1983; Potter 1991; Tobón
2004). Numerous studies reported that fluxes of dissolved nutrients are highly correlated with
the amount, duration and the intensity of precipitation penetrating the canopy (Reiners 1984;
Parker 1983; Lovett 1996; Tobón 2004; Potter 1991) that are much higher in tropical forests
than in temperate forests. A high intensity of rainfall leads to a decrease in ion concentration in
TF during the event while low intensities correspond to elevated solute concentrations.
Precipitation of low intensity prolongs the residence time for water within the canopy and hence
exchange processes have more time to take place (Lovett 1992; Hansen 1994). Throughfall
chemistry also strongly depends on the type of precipitation such as clouds, droplets, mist and
dew. Indirect horizontal precipitation such as clouds have elevated concentrations of elements
and can be important especially in tropical montane cloud forests where quantities of
throughfall can exceed those of bulk precipitation and therefore cause an extra input of water
(Cavelier 1997) and nutrients (Veneklaas 1990).
Factors such as altitude, forest type and canopy structure, stand age or successional state,
species composition, seasonality and nutrient supply are also important factors influencing
throughfall fluxes. The leaf area index (LAI) as an index for the canopy density also influences
the amount of throughfall or rather interception loss. During the drier season in a tropical
montane cloud forest the interception by the canopy was lower than during the rainy season
probably due to a decreased LAI because of less growth and an interception maximum at the
beginning of the wet season as a result of increased LAI from leaf flush (Cavelier 1997). Potter
(1991) reported also up to 12 times higher NTF during the growing season than during
dormant season in the Appalachian forests, indicating the considerable influence of foliage in
altering throughfall chemistry. The occurrence of epiphytes may increase also wet and dry
deposition through the enlargement of the surface area for impaction and sedimentation
(Reiners 1984).
Moreover Hansen (1994) reported that ion concentrations in TF during the initial part of a rain
event (the first 2 mm) were highest after long dry periods which indicated that inter-event
deposited (external source) and probably foliar material (internal source) are easily leached in
the early phase of wetting. In subsequent rain the ion concentrations generally decrease except
for NH4+ and K+ which showed a high variability. Atmospheric nutrient inputs (aerosols, dust or
gases) also vary with proximity to sources such as the ocean, volcanoes, industries and
agriculture (Attiwill 1993; Parker 1983). High concentrations of Na+, Cl- and Mg2+ are often
closely related to the sea nearby whereas enhanced N, P and S concentrations are linked to
deposition of air pollutants and hence derive from anthropogenic sources.
- 7 -
Nevertheless, canopy exchange processes are also related to the chemical composition in
incident precipitation. If the concentration of a certain nutrient is high, net fluxes are more
likely to be negative i.e. a net uptake by the canopy while low concentrations in rain lead to a
positive net flux from the canopy to the forest soil i.e. a net loss of nutrients (Veneklaas 1990).
Highly leachable elements, e.g. potassium, do not follow this pattern.
As mentioned above the nature of the canopy such as epiphytic colonization and abundance of
lichens and bryophytes affect the interception and alteration of rainfall. The greater the canopy
roughness which means more species with different crown structures and more gaps in the
upper canopy, the higher dry deposition rates may occur and the more precipitation is
intercepted by the canopy (Lovett 1984; Veneklaas 1990). Forti (1991) found out that ion
concentrations of wet deposition samples above the canopy compared to those from clearing
are overestimated due to the fact that the roughness of the upper canopy layer generates
vortices and turbulent mass flow at the top of the canopy which resulted in enrichment of
aerosols above the canopy and an increase of ions in incident precipitation. In contrast, Hansen
(1996) reported that due to these turbulent air movements less incident precipitation was
collected and that ion concentrations above the canopy and from an open field were similar.
Dry deposition and canopy exchange as well are expected to change with canopy depth
(Hansen 1994; Parker 1983; Lovett 1992). DD is generally assumed to be highest in the upper
canopy due to their greater exposure to wind and turbulent air whereas in lower canopy layers
the incident rainfall is already enriched in nutrients through their wash-off or leaching from the
upper layers.
Evaporation of intercepted rainwater in inter-event periods can cause an increase of ion
concentration on foliage surfaces previously dissolved in rainwater. Additionally
evapotranspiration exceeds incident rainfall in dry seasons and may provoke an increase in
transpiration and thereby in nutrient release from leaves through their stomata (Parker 1983).
The main processes affecting the variability in throughfall chemistry, dry deposition and canopy
exchange (Tobón 2004; Ragsdale 1992; Lovett 1984), will be discussed in further detail below.
1.4. Dry deposition
Atmospheric deposition
Atmospheric deposition represents an external input of elements or nutrients to ecosystems
and includes wet deposition and dry deposition. Wet deposition comprises incident precipitation
i.e. water droplets in which aerosols, gases or ions are dissolved. Solutes can originate from
other locations nearby or distant (e.g. the ocean) or dissolution took place shortly before
- 8 -
reaching the top of the canopy. Dry deposition combines all input processes between rain
events and is made up of physical processes such as sedimentation and impaction of particles
and aerosols and absorption of gaseous forms (Parker 1983). An additional crucial biotic
nutrient input within the canopy is the fixation of atmospheric nitrogen (N2) by free-living
cyanobacteria or symbiotic epiphyllous microorganisms (lichens) growing on leaf surfaces. A
considerable portion of this newly fixed N by epiphylls may be transferred to the host leaf
accounting for 10-25% of total leaf N content. Moreover N2 fixation rates are highly correlated
with forest and foliage age due to the time needed for colonization of epiphylls on leaf surfaces
(Bentley 1987; Bentley 1984).
Dry deposition
Dry deposited materials originate from natural sources such as soil processes through microbial
activity and anthropogenic sources: combustion of fossil fuel and coal, biomass burning,
industrial plants, agriculture and animal waste (Lovett 1992). Particulate dry deposits are
classified in dust, sand, volcanic ash, sea salts, air pollutants and fertilizer and related ions are
Ca2+, Mg2+, Na+, Cl-, NO3-, NH4
+, SO42-, PO4
3-. Human activities are also responsible for notable
emissions of gases and aerosols such as SO2, NOx (NO and NO2), NH3 and HNO3. Furthermore,
Veneklaas (1990) and Kellman (1982) reported sporadic nutrient inputs to tropical ecosystems
through proximate volcanic activity.
Despite N2 being the most abundant single element in the atmosphere with about 78% by
volume and its key role in plant nutrition, it is not available for higher plants since it occurs in
the non-reactive form N2. As mentioned above some bacteria or algae can fix N2 and
incorporate it into organic matter though its contribution to total N fluxes within a forest-
ecosystem is generally considered to be low. The less abundant but reactive N forms, mainly
oxides (NO, NO2, N2O, HNO3 vapour, NO3-) but also reduced forms (NH4
+ and NH3), are crucial
in atmospheric chemical reactions. Dry deposition of N primarily occurs through gaseous HNO3
and NH3 because they are more reactive and are easily absorbed by the plant surfaces (Lovett
1992; Hill 2005). Moreover, Balestrini (2001) found for alpine forests a close correlation
between nitrate/ sulphate and ammonium/ hydrogen deposition and presumed that these ions
were mostly deposited as ammonium sulphate and nitrate aerosols resulting from ammonia
and sulphuric and nitric acid reactions. At tropical latidues land use change e.g. conversion of
forest in agricultural lands and increasing biomass burning are assumed to be responsible for
increased N emissions to the atmosphere.
A large fraction (40-50%) of deposited inorganic N was captured by the canopy of temperate
forests because of uptake by plants and epiphytic lichens and of microbial immobilization
- 9 -
(Garten 1998). This basically agrees with investigations from Clark (1998) who reported that
80% of NO3--N and 61% of NH4
+-N from atmospheric N deposition were retained by the canopy
in a tropical montane rain forest and in turn epiphytic bryophytes and its aggregations with
vascular epiphytes accounted for 80% of this retention. Experiments from Wilson (1998)
showed that uptake of wet deposited inorganic N made a small but sizeable contribution to the
total N demand of the vegetation. Moreover Garten (1998) suggested that the canopy uptake
of dry deposited N is a major contributor to N fluxes within the canopy. Hence throughfall
measurements underestimate the atmospheric deposition of N compounds to a large degree
(Hansen 1996).
The dry deposited material within or on the top of the canopy can enter the nutrient cycle of
the forest-ecosystem by several pathways: washing off of particulate substances and aerosols
from foliage surfaces by incident precipitation or absorption of gases through stomata and
cuticlefree spaces on tissue surfaces (ectodesmata) as well dissolution in the water film on the
leaf surface. The dissolved material then can either drip down as throughfall to the forest floor
or can be taken up by the leaves or associated epiphytes and epiphylls by ion exchange
processes. Dry deposition rates are mainly influenced by the duration of inter-event periods
and further by wind speed and turbulence, deposition velocity and the canopy structure such as
foliage characteristics, surface roughness and wetness (Lovett 1992).
1.5. Canopy exchange processes
In the last decades numerous studies paid attention to canopy exchange processes, focusing on
the extent to which they contribute to solute fluxes from the atmosphere to the soil surface
commun.). Mean annual temperature between 1997 and 2001 was 27.4°C (monthly
average from 23.2°C to 31.5°C) at the Tropical Field Station La Gamba, which is next to the
Esquinas Rainforest, and 25.2°C (monthly average from 22.3°C to 28°C) inside the forest
(Weissenhofer and Huber 2001). According to Pamperl (2001) the prevailing soils are, in the
order of occurrence, ulitsols, inceptisols and entisols.
The experiments were conducted in an undisturbed primary ravine forest and a secondary
ravine forest (age 25 yrs) the latter one has been used as a cacao plantation.
Tracer techniques
Preliminary tests were carried out in February 2005; the main experiments were conducted
in September and October 2005. For the tracer experiments two different labelling
techniques were developed and performed: (1) stem injection of 15N, Rb+ and Sr2+ for
subcanopy trees (5-12 m height) along with a controlled leaf surface wash experiment and
(2) application of 15N-enriched simulated rain events (spraying on the canopy) to study
gross and net nitrogen fluxes on selected host tree branches that are densely covered with
vascular and cryptogamic epiphytes.
- 25 -
(1) Stem injection of 15N, Rb+ and Sr2+
Plant material and experimental setup
For the stem injection approach appropriate understorey tree species with easily reachable
and adequately developed crown, low or no epiphytic colonization and a sufficient number of
plans were selected at the two forest sites, namely Psychotria elata (Sw) Hammel
(Rubiaceae) in the primary ravine forest and Siparuna thecaphora (Poepp. & Endl.) A. DC.
(Monimiaceae) in the secondary ravine forest. Four trees of S. thecaphora were treated with 15N- and Rb+/Sr2+-solution and each four trees of P. elata were labelled either with 15N- or
Rb+/Sr2+-solution.
To inject the tracer solutions into the xylem channels stems were drilled with a borer (inner
diameter = 4 mm) to about 3 mm depth and a short metal tube was fitted into the hole.
The tube was connected by a flexible PVC tube to a 20 ml syringe reservoir (plastic) which
was attached above the hole. The reservoir and connecting tubes were filled with 20 mL
labelling solution (bubble free) and the reservoir was covered with parafilm and fixed to the
plant stem with tape. The 15N-labelling solution was made of Na15NO3, 15NH4Cl and
15NH415NO3 (equimolar) with a total N concentration of 60 mM and 98 at % 15N and the
Rb+/Sr2+-solution consisted of RbCl (100 mM) and SrCl2 (50 mM). Rb+ has been shown to
behave (bio-) chemically similar to K+ (Nyholm and Tyler 2000) and Rb-radioisotopes have
e.g. been used to study K+ uptake kinetics and fluxes in roots. Ca2+ and Sr2+ have been
applied to infer ecosystem Ca2+ sources and dynamics. We therefore used rubidium as
tracer for potassium and strontium as tracer for calcium fluxes within the canopy of
understorey trees. During the treatment period (two weeks) the solution was replenished
and the amount taken up by the plant was recorded daily.
Sampling
Sampling of rain water above and below the treated plants started after two weeks of tracer
amendment and included 3 rain events (< 10 mm to minimize dilution of ion concentrations
in throughfall) for each plant in short intervals. Water samples (throughfall above the
canopy and below the canopy) were collected in four replicates (plants), i.e. four trees
labelled by stem injection, per rain event. Each throughfall sampler was constructed from
two gutters (0.75 m length) fixed above and beneath the canopy of the sample tree on a
bamboo-pillar. Both gutters were provided with an outlet where the incident rainfall was
channelled through flexible PVC tubes down into plastic flasks (1.5 L). The flasks were
emptied, cleaned carefully and washed with deionized water after each rain event during
the sampling period. Aliquots of bulk precipitation samples from the open field at the Field
Station La Gamba were also taken. The collected rain water (approx. 100 mL each) was
poured into PE flasks and HgCl2 was added (0.1 mM final concentration) to avoid microbial
growth and alteration of nutrient content in throughfall. The water samples were kept frozen
until transport to Vienna.
- 26 -
Leaf samples of each tree were collected before starting the labelling and after the labelling
period, between the two throughfall samplers after every collected rain event. One half of
fresh leaves was treated in a microwave oven to stop enzymatic activities and further dried
at 60 °C for 72 h. The other half of labelled leaf samples was used for controlled leaching
experiments. For this three to five leaves of each leaf sample per plant and event were put
into a plastic bag with zip and a defined amount (100 mL to 200 mL, depending on leaf
size) of deionized water (>18.2 MΩ cm-1) was added. After two hours the leaching
experiment was stopped and water with leachates was collected in scintillation flasks and
treated similar as throughfall samples. The fresh leached leaves were dried superficially,
scanned to determine leaf area and finally treated in the microwave oven and dried as
described above.
(2) Application of 15N-enriched simulated rain events
Plant material and experimental setup
For the 15N-enriched rainwater application branches (1.0-1.5 m length) with a typical load of
vascular epiphytes, epiphytic bryophytes and epiphylls on host leaves from the primary
ravine forest were selected. Two horizontal branches of Alchornea cf.costaricensis Pax & K.
Hoffm. (Euphorbiaceae) and three horizontal branches of Ficus tonduzii Standl. (Moraceae)
were cut and hung horizontally under a tin roof at the forest edge in the garden of the
Tropical Field Station La Gamba to avoid uncontrolled contact with incident rainfall. The cut
surface of the host branch was covered with water-soaked tissue in a plastic bag taped to
the branch to minimize water stress of host leaves.
Rainwater was collected on an event basis in a big PVC barrel (200 L) in the open field, each
before a simulated rain event as far as it was raining. In intervening periods with no rainfall
the barrel was covered by a plastic foil and put into the shadow to minimize chemical
alteration. The first simulated rain event applied was enriched with the stable isotope 15N in
equimolar amounts of NO3-, NH4
+ and DON (0.55 µM 15NH415NO3/
15N-L-glutamic acid
equivalent to 10 at% 15N). The following simulated rain events were conducted with
unlabelled rainwater. Under each branch a plastic cover was fixed to ensure complete
throughfall collection. The dripped off throughfall was channelled and filtered through small
mesh nylon into polyethylene canisters. Before starting the experiment the branches were
sprayed with unlabelled rainwater to test water retention and the amount of water required
to penetrate the whole branches and its canopy components. The amount was depending on
the load size of epiphytes (1.5-2.5 L). To provide enough time for solute exchange every
single simulated rain event was applied in two aliquots of 6 to 14 min duration with an
interval of about one hour in between. The branches were sprayed evenly with a defined
amount of rainwater using a pressure sprayer to simulate drizzle. Simulated rain events
were conducted twice a day in constant intervals over a period of four to five days. This was
- 27 -
the maximal possible time before the host leaves started to loose turgor. Separate
equipment was used for 15N labelled rain event and normal rain events, respectively, and
the equipment was washed and cleaned carefully after each rain event.
Sampling
To quantify the amount of rainwater applied the remaining volume within the pressure
sprayer was weighed. After a drip off time (two hours) the volume of throughfall collected in
the polyethylene canisters was measured. Aliquots (250 mL) of rainwater and throughfall
filtered through fine mesh nylon were transferred into PE bottles after each rain event,
HgCl2 was added (0.1 mM final concentration) and samples were kept frozen until transport.
At the end of the experiment the branches were divided into different canopy components:
soil, canopy litter, bark and wood. Unlabelled reference material of canopy components and
host leaves were sampled from other branches of the same trees at the original place. The
fractions were treated in a microwave to stop microbial or enzymatic activities and the
material further dried at 60 °C for 72 h. Dry weight of each fraction was determined and
samples were weighed and packed airtight until transport to Vienna.
15N-Pool-Dilution approach
Pool dilution techniques are commonly used to estimate gross N transformation rates in
soils (Luxhoi 2004). For instance to measure gross N mineralization the NH4+-pool is
labelled by addition of 15NH4+. N mineralization produces NH4
+ from an unlabelled organic N
pool, therefore adds 14NH4+ to the pool and ultimately leads to dilution of the 15N:14N signal
in the NH4+-pool. In contrast, processes that consume NH4
+ such as microbial uptake
(immobilization), plant uptake or nitrification, do not change the 15N:14N signature of this
pool and utilize NH4+ at the 15N:14N ratio present in the NH4
+-pool.
This method was transferred to gross canopy fluxes of N. The N fractions (NO3-, NH4
+ and
DON) of rainwater before canopy contact represent the pools which are labelled with 15N.
During penetration through and interception by the canopy the pools are altered. Inputs of 14N may occur through canopy leaching or N2 fixation processes and 15N:14N losses may
happen due to canopy uptake processes. Determination of isotope signatures and N
concentrations enabled the quantification of gross and net N canopy fluxes. Furthermore, 15N dilution is studied along a time axis of e.g. 4 to 24 hours. In the approach transferred to
canopy processes the time axis is represented by rainwater before and after canopy
exchange and rates are therefore presented on an event basis and not per hour or day.
- 28 -
Fig.1: Scheme of the 15N-Pool-Dilution approach for (A) determination of gross N mineralization rates and (B)
transferred to gross canopy fluxes of N; WD: wet deposition, TF: throughfall
Chemical analyses
Chemical analyses were carried out in Vienna, Department of Chemical Ecology and
Ecosystem Research, University of Vienna.
Stable isotope analyses
Dried organic materials such as leaves, bark and epiphytes of both experiments were
redried at 80°C for 24h and ground to a fine powder in a ball mill (Retsch MM2 and MM200).
Aliquots of 1-2 mg were weighed into Sn-capsules and 15N and total N content were
analysed by continuous-flow IRMS (isotope ratio mass spectrometer). The system was
made up of an elemental analyser (EA 1110, CE Instruments, Milan, Italy) which was
connected to the IRMS (DeltaPLUS, Finnigan MAT, Bremen, Germany). Reference gas (high
purity N2, AGA, Vienna, Austria) was calibrated to the atmospheric N2 standard (at-air)
using IAEA-NO3, IAEA-N-1, and -2 reference material (International Atomic Energy Agency,
Vienna, Austria).
Aliquots (50 mL) of rain samples (throughfall above and below the canopy, and leachates)
of the stem injection experiment were therefore evaporated to dryness with a rotary
evaporator (R-114/ R-124, Büchi, Switzerland). Samples were redissolved in 0.5 mL
distilled water, dried again using a SpeedVac concentrator (SC 110, Savant, NY) and
redissolved in 100 µL distilled water. Aliquots (50 µL) were dried in Sn-capsules and
analysed for 15N and total N content by IRMS.
Water samples of the simulated rain application tracer experiment
For stable isotope analyses rain and throughfall samples of the 15N-enriched rain event (i.e. 15N-Pool dilution approach) and the consecutive two unlabelled rain events were used. Rain
and throughfall samples of the 15N-enriched rain event and the first unlabelled rain event
Mineralization
Consumption
+ 15N addition
+ 14N
NH4+ pool
in soil
- 15N:14N
15N:14NH4+
(A)
Canopy Leaching
Canopy Uptake
+ 15N addition
+ 14N
RainwaterWD & TF
- 15N:14N
15N:14NH4+
15N:14NO3-
15N:14DON
(B)
- 29 -
were fractionated into the N forms ammonium, nitrate and dissolved organic nitrogen
(DON). Water samples of the second unlabelled rain event were treated as mentioned
above. Small aliquots of original samples were analysed for anions and cations by HPLC (DX
500, Dionex, Vienna, Austria) and total dissolved N and DOC concentration by NPOC/TDN-
analyzer (Shimadzu, Japan) as well.
Isolation of N fractions
According to Hertenberger and Wanek (2004) the microdiffusion approach to isolate
ammonium can be combined with common cation-exchange chromatography to separate
the organic nitrogen fraction from nitrate. Sample volumes were determined, pH was
adjusted to 5-6 with 0.1 N HCL to avoid volatile loss of NH3 or NO2 during evaporation, and
then samples were concentrated under reduced pressure to about 10 to 15 ml with a rotary
evaporator and weighed again. The concentrates were poured into Schott-flasks (50 mL)
and acid traps were added. Acid traps were made of quartz fibre filter discs containing 12 µL
2N H2SO4 which were enclosed in a strip of Teflon band that is gas permeable but
waterproof. After addition of MgO (100 mg) to increase pH >9.0, the bottles were
immediately closed and NH3 released into the head space was captured by the floating acid
traps. The flasks were placed on a shaker for 5-7 days (75 rpm, 37°C). Then acid traps, still
enclosed in Teflon band, were removed and dried over conc. H2SO4 under vacuum
(24-48 h). The dry filter discs were transferred into Sn-capsules shortly before IRMS-
analysis to avoid corrosion of the capsules. The remaining concentrates containing nitrate
and DON were filtered to remove MgO and pH was adjusted to 5-6. Samples were then
H+-form, Fluka). The columns were washed two times with 10 mL distilled water to collect
nitrate in the flow-through and the eluate was neutralized with 0.1 N KOH. The nitrate
fractions were evaporated to dryness redissolved in 0.5 mL distilled water, again dried by
SpeedVac and redissolved in 70 µL distilled water. Samples were quantitatively transferred
and dried in Sn-capsules and analysed for 15N and total N content by IRMS. The adsorbed
DON fractions were eluted from the resin with 25 mL 1 N HCL and treated as the nitrate
fractions after addition of 100 µL 1 N KOH.
Ion, NPOC and TDN analyses
Hot-water extracts were prepared by extracting aliquots of 20 mg ground leaf material with
1 mL deionized water at 95°C for 60 min. Inorganic cations (K+, Rb+, NH4+, Na+, Ca2+, Sr2+,
Mg2+) and anions (Cl-, NO3-, SO4
2-, PO43-) in rain samples (bulk precipitation, throughfall
above and below the canopy, and leachates) and hot-water extracts of the stem injection
experiment were analysed by HPLC (high pressure liquid chromatography, DX 500, Dionex,
Vienna, Austria) and conductivity detection. Anions were separated on an anion exchange
column (AS11, 4x250 mm, Dionex) using a linear KOH gradient (2 to 40 mM in 6 min, total
run time 12 min). Cations were separated on a cation exchange column (CS16, 5x250 mm,
- 30 -
Dionex) by an isocratic method with methanesulfonic acid as eluent (28 mM, 45 min,
60 °C). Rain water samples of the simulated rain events were analysed for anions (Cl-, NO3-,
SO42-, PO4
3-) as mentioned above and but method to analyse cations (K+, NH4+, Na+, Ca2+,
Mg2+) were modified (30 mM methanesulfonic acid for 26 min and 40 °C).
NPOC (non purchable organic carbon) and TDN (total dissolved nitrogen) of rain samples
(bulk precipitation, throughfall above and below the canopy from both experiments) were
determined using a TOC-VCPH/CPN/ TNM-1 analyzer (Shimadzu, Japan). Inorganic carbon was
automatically removed during measurement through addition of 2 N HCl and purging with
synthetic air (CO2 free). DON was calculated by subtracting DIN (dissolved nitrate and
ammonium) from measured TN.
Calculations
Enrichment factor for the stem injection experiment
The enrichment factor (EF) for TFb (throughfall below the experimental plant) was
calculated as:
(i) EF = c(TFb)/ c(TFa)
where c is the concentration of the respective solute in TFb and TFa (throughfall above the
experimental plant)
Two-source mixing model for the stem injection experiment
The fractional contribution of leaching from host tree foliage to throughfall below the canopy
(TFb) was estimated for potassium, calcium and nitrogen with a modified two-source mixing
model:
(ii) % PLEA = 100 x (RTFa - RTFb)/ (RTFa – RLEA)
where PLEA represents the relative fraction of potassium, calcium or nitrogen, leached from
foliage of labelled woody plants to throughfall below the canopy (TFb), R the ratios of
tracers e.g. Rb+:K+, Sr2+:Ca2+ or 15N:14N in TFa, TFb, and leachates from controlled
leaching-experiment (LEA). In consideration of the dynamic model CE (canopy exchange)
and DD (dry deposition) are estimated in absolute numbers:
(iii) CE = TFb x % PLEA = TFa x EF x % PLEA
For TFa an arbitrary concentration of 1 is assumed and CE represents the relative
contribution of canopy exchange to TFb. Consequential the relative contribution of DD to
TFb can be estimated as:
- 31 -
(iv) DD = TFb – (TFa + CE)
15N-Pool-Dilution equation
The equations to determine gross N mineralization or nitrification from Bengston (2006)
were adopted and modified to calculate gross N canopy fluxes:
ln ((fTF-k)/(fWD-k))ln (WTF/WWD)
(v) Efflux = x WWD - WTF
ln ((fTF-k)/(fWD-k))ln (WTF/WWD)
(1 + ) x WWD - WTF(vi) Influx =
where k represents the natural abundance of 15N in rainwater (at% 15Nunlabelled WD), (fWD-k)
the APE (atom percent excess) of 15N in applied rainwater (wet deposition, WD), (fTF-k) the
APE of 15N in throughfall (TF), WWD the initial N pool in wet deposition (in µg N), and WTF the
N pool in TF (in µg N) of the respective fractions i.e. ammonium, nitrate or DON. Therefore,
each flux was calculated separately for NH4+, NO3
- and DON.
Statistical analyses
Statistical analysis was performed with STATGRAPHICS PLUS 5.0. Software (Statistical
Graphics Corp., Rockville, MD, USA). Differences between three or more groups were tested
by one-way analysis of variance (ANOVA) (multiple range test, LSD test) (P <0.05).
RESULTS
Stem injection experiment
The stem injection experiment was performed to examine the potential contribution of
subcanopy trees to nutrient enrichments in throughfall through foliage leaching. In
preliminary tests in February 2007 we observed that stem injection was an effective
labelling method for the understorey trees P. elata (representative for secondary forest) and
S. thecaphora (representative for primary forest). Foliar, hot-water extractable
concentrations of rubidium and strontium increased 15-fold and 1.5-fold, respectively,
compared to unlabelled plants, and δ15N values ranged from 80 to 2000 (data not shown).
During the main experiment conducted in September and October 2007 a
considerable enrichment of rubidium, strontium, and 15N in leaves was reached by stem
injection (Fig.2, closed bars). Moreover controlled leaching experiments were carried out to
determine tracer leaching efficiency of labelled leaves. Similar ratios were detected for
- 32 -
Rb+:K+ and Sr2+:Ca2+ in leachates (LEA) compared to leaves (L) in S. thecaphora (Fig. 2D
and 2E). P. elata exhibited a markedly lower leaching efficiency for all tracers applied, as
also shown for 15N in S. thecaphora (Fig 2A, 2B, 2C, and 2F).
Ratios of tracers and their corresponding elements (Rb+:K+, Sr2+:Ca2+ and 15N:14N) in bulk
precipitation, collected in the open field (P), were slightly higher in comparison to those in
throughfall collected above the canopy of treated plants (TFa), though no significant
differences were observed.
Comparison of tracer:element ratios of TFa and TFb (throughfall below the canopy) showed
no statistically significant differences. Enrichment factors for solutes in throughfall (TFb) are
shown in Table 1. Concentration of all determined solutes were enriched in TFb with
greatest enrichment factors in PO43- (3.17) followed by K+ (2.52) and the other solutes
(1.18 - 1.96). Interestingly EF was lowest for all N solutes, Na+ and DOC (Table 1).
Based on concentration ratios of Rb+:K+, and Sr2+:Ca2+, respectively, and on a modified
two-source mixing model we estimated the fractional contribution of leaf leachates from P.
elata and S. thecaphora to throughfall (TFb) enrichment of K+ and Ca2+ (Fig. 2A, 2D, and
2E). The two sources of K+ (Rb+) and Ca2+ (Sr2+) to TFb are TFa and LEA. This approach
was only used where the two sources, TFa and LEA were significantly different i.e. Rb+:K+ in
P. elata and S. thecaphora and Sr2+:Ca2+ in S. thecaphora (Fig. 2A, 2D, and 2E). We
assumed that TFb was composed of TFa and LEA and from dry deposition (DD). The ratios
of Rb+:K+ and Sr2+:Ca2+, respectively, in LEA represented the maximum possible
contribution of foliage leaching of woody plants to TFb, while the respective ratios in TFa
represent the minimum contribution by LEA. The results showed that S. thecaphora
considerably contributed to throughfall enrichment of both potassium and calcium via
foliage leaching i.e. 14% and 32%, respectively. P. elata contributed 32% of K+ to TFb by
leaching.
Though 15N:14N ratios were significantly higher in bulk leaves than in all other fractions the
tracer (15N) did not show up in leachable pools at the time of experiment (Fig. 2C, 2F). The 15N:14N ratios therefore did not differ significantly between LEA, P, TFa, and TFb for both P.
elata and S. thecaphora and therefore hindered the application of the two source model to
evaluate N sources in throughfall (see also Sr2+:Ca2+ in P. elata, Fig, 2B).
Simulated rain experiment
Net fluxes
A total of five branches (A. cf. costaricensis plus F. tonduzzi) where treated with simulated
rain events twice a day over a period of 3 days where the first applied event was enriched
with the stable isotope 15N. Net fluxes were calculated as volume-weighted concentrations
of TF minus those of WD (wet deposition). Negative net fluxes indicate net leaching of
- 33 -
canopy components while positive net fluxes refer to net uptake. Time dependent course of
event based net fluxes of N fractions (NO3-, NH4
+, DON – dissolved organic nitrogen), base
cations (K+, Mg2+, Ca2+ and Na+), anions (SO42-, PO4
3-) and DOC (dissolved organic carbon)
showed a high variability throughout the experiment but no correlation among ions and
fractions were evident, respectively (data not shown). Thus we calculated the sum of net
fluxes of seven events conducted within three days (Fig. 3). To increase the comparability
between branches event based and integrated net fluxes were calculated on the basis of the
sum of dry weights of potentially canopy components contributing to net solute exchange
The recovery of 15N, applied by the first simulated rain event, in throughfall and canopy
components, amounted to 69.8%. Throughfall, the first three events were measured for 15N, accounted for almost two thirds of the 15N detected (43.3%). About 96% of throughfall 15N was found in throughfall collected during the first 15N-enriched rain event. The
partitioning of 15N within the canopy is shown in Figure 5B in percent of initial 15N amount
applied in rainwater. The highest quantity of incorporated 15N was determined in epiphytic
bryophytes with 16.4% or 0.2 µmol 15N. All other canopy fractions did not differ significantly
although epiphylls and leaves exhibited somewhat higher 15N contents (3.3% and 3.2%,
respectively) than residual fractions whose proportions ranged from 1.3% to 0.4% in the
descending order: bark, canopy litter, canopy soil, vascular epiphytes and lichens. 15N uptake rates of canopy components were estimated (Fig. 5C) but in contrast to
distribution of 15N within the biomass, the epiphyll-fraction showed the highest uptake rates
lichens and mosses incorporated 15N from incident rainwater at a lower rate (0.016 ± 0.008
to 0.012 ± 0.005 µmol 15N g-1 DW) compared to epiphylls. Despite leaves made up the
highest proportion of dry mass their uptake rates were lowest (0.001 µmol 15N g-1 DW).
DISCUSSION
Models to demonstrate controls of throughfall chemistry
We developed two models to interpret alterations in throughfall chemistry and their control
mechanisms (Fig. 6). On the one hand we adopted a static model which is based on
unidirectional net fluxes (NTF) alone (Fig. 6, left plot) which can be either negative or
positive, or zero. A negative NTF of a solute corresponds to a decrease in its concentration
- 35 -
after penetration of the canopy (Fig. 6, A) and is generally caused through uptake
processes within the canopy. By contrast an increase in nutrient concentrations in
throughfall results in positive net throughfall fluxes and is caused by foliar leaching and/ or
dry deposition (Fig. 6, C). Apart from nutrient depletions or enrichments in throughfall
solute fluxes of incident rainfall (TFa) and throughfall (TFb) may be balanced and the
resulting net flux is zero (Fig. 6, B) although exchange processes take place within the
canopy. In this case, investigations of net nutrient fluxes are inadequate to explain the
nutrient dynamics within the canopy. Canopy exchange in situations of zero net fluxes can
only be demonstrated by tracer studies.
Thus, we additionally put forward a dynamic model (Fig. 6, right plot) which implies
alterations in throughfall chemistry and concentrations through bidirectional fluxes i.e.
foliar leaching and foliar uptake. For instance, the concentration of a solute may not
change during penetration of the canopy, therefore the net nutrient flux is zero. However,
the solute might be taken up by some canopy components and be released by others at
the same time. According to the extent of canopy exchange fluxes, the dynamic model is
illustrated by three different scenarios: scenario one (Fig. 6, sc1) indicates no contribution
of exchange processes to throughfall chemistry, in scenario two (Fig. 6, sc2) 50% of
solutes are exchanged and scenario three (Fig. 6, sc3) shows a complete exchange of a
nutrient in rainwater during canopy passage.
To investigate these exchange processes and given that net fluxes do not provide answers
to the origin of soluble nutrients in throughfall, we conducted tracer experiments for
selected nutrients, namely potassium, calcium, and nitrogen, to follow their pathways
within the canopy. Tracers used were Rb+, Sr2+, and 15N that behave (bio-) chemically
similar to the tracee. Stem injection tracer experiments along with controlled foliar
leaching experiments, may allow differentiating whether nutrient enrichment in throughfall
resulted from dry deposition or foliage exchange and to what extent (Draaijers et al.
1997).
We demonstrate four cases in the stem injection experiment (Fig. 7) where tracer:element
ratios of throughfall (TFb) are compared to the corresponding ratios of throughfall above
the labelled canopy (TFa) and foliar leachates (LEA). Case 1 does not allow any derivation
of the fractional contribution of dry deposition or foliage exchange to TFb enrichment since
no significant amounts of tracer were found in leachates. The remaining cases (Fig. 7, case
2-4) allow detecting potential foliage exchange effects on throughfall chemistry. In case 2
ratios of TFb and TFa do not differ although the tracer was found in LEA. This suggests that
foliar exchange does not contribute to TFb chemistry. Net positive fluxes of potassium,
calcium or nitrogen in throughfall therefore are controlled by dry deposition and for
nitrogen by N2-fixation. Further this case agrees with scenario one from the dynamic
model. In case 3, TFb is obviously tracer enriched (higher ratio) due to foliar exchange
- 36 -
contributing about half to TFa but the contribution of dry deposition can not be estimated
with this approach and fits with scenario 2 of the dynamic model. The model can be
expanded based on enrichment factors (EF) and tracer:element ratios to derive the fraction
of canopy exchange (CE) and dry deposition (DD) (see (i) and (ii)). Finally in case 4,
according to scenario three of the dynamic model, throughfall chemistry is replaced
completely by foliar uptake and -leaching.
Throughfall enrichment
Measurements of throughfall chemistry below the canopy of understorey trees (P. elata
and S. thecaphora) in a tropical lowland forest showed an increase of all examined solutes
and therefore positive net throughfall fluxes (Table 1). We hypothesized that TF
enrichment of potassium was controlled by foliar leaching, whereas calcium derived mainly
from dry deposition and nitrogen originated from both foliar leaching and dry deposition,
and from N2-fixation.
In three cases of the stem injection experiment (Fig. 2A, 2D and 2E) injected rubidium and
strontium was found in leachates therefore allowing to estimate the contribution of canopy
exchange (leaching) to TFb for potassium and calcium in the understorey tree S.
thecaphora (in the primary ravine forest) and for potassium in P. elata (in the secondary
forest). According to the static model the relative contribution of foliar leaching to
potassium and calcium in throughfall below S. thecaphora canopies was 14% and 32%,
respectively, and for potassium in throughfall under P. elata 32% (see equation (ii)).
Considering that net leaching of solutes to throughfall is the result of canopy exchange
processes their contribution was evaluated by the dynamic model (see equation (iii) and
(iv)). K+ in throughfall under P. elata seems to derive to almost equivalent parts from foliar
exchange (4/9) and dry deposition (5/9) whereas potassium and calcium concentrations in
throughfall under S. thecaphora were controlled by canopy exchange (3/4). In the case of
P. elata results of the tracer experiment are contradicting to our hypothesis as it is
generally agreed that potassium derives mainly from canopy leaching due to its extreme
mobility and high apoplastic K+ concentration in plant tissues (Parker 1983; Sattelmacher
2001). Comparative data for dry deposition fluxes of potassium were found for temperate
forests and tropical rain forests. For a mixed hardwood forest (Lindberg et al. 1986)
estimated that dry deposition contributed about 60% of potassium as coarse particles.
Lovett and Lindberg (1984) suggested that dry deposited K+ derived from suspended soil
or biological material. Moreover they observed highest K+-aerosol concentrations from
within the forest itself indicating an in-canopy source.
Calcium in throughfall is considered to almost equally originate from wash off of particulate
dry deposition (Lovett and Lindberg 1984; Lindberg et al. 1986; Forti and Neal 1992) and
foliar leaching (Tukey 1970; Parker 1983). However, we observed that a large portion
- 37 -
derives from foliage leaching. Calcium is generally tightly bound in structural plant tissue,
e.g. cell walls, and leaching of calcium mostly occurs from exchangeable pools in leaves,
where calcium has been recently accumulated. Moreover, mature tissues are more
susceptible to leaching than younger leaves since the latter accumulate calcium directly
within cell walls and have a smaller exchangeable pool for calcium (Mecklenburg et al.
1966). Though Hansen et al. (1994 and 1996) reported that dry deposition rather has an
impact on throughfall fluxes in the upper canopy layers, where trees are highly exposed to
turbulent air movements, our results indicate an effect of dry deposition on K+ and Ca2+ in
the understorey.
In the other three cases the stem injection experiment failed, including those for nitrogen
for both species (Fig. 2B, 2C and 2F). We therefore could not clarify whether increased
concentrations of nitrogen and calcium in TFb were caused by dry deposition, foliar
exchange or in the case of nitrogen by atmospheric nitrogen fixation as well. Although the
tracer injection was successful and tracer was allocated to foliage, this signal was not
found in foliar leachates. It is therefore likely that the injected tracers, 15N and strontium,
were incorporated and immobilized in stable pools within the plant, e.g. cell walls and
proteins, and therefore were removed from the exchangeable pool, the apoplast solution,
during the two weeks of injection.
Consequently this experiment showed that foliar leaching from understorey woody plants
contributes to throughfall enrichment of potassium and calcium by 40-75% in a lowland
tropical rain forest. Nevertheless, the experimental setup was not suitable to investigate N
exchange processes within the canopy and to determine whether canopy impacts are
controlled by unidirectional (static model) or bidirectional (exchange) fluxes (dynamic
model). Though we were able to estimate net canopy fluxes (DD, leaching) the underlying
gross fluxes i.e. canopy uptake and canopy leaching, could not be studied by this
approach. Therefore we adopted the 15N isotopic pool dilution approach, generally used to
determine gross N mineralization rates in soils (Bengston 2006), to detect gross canopy N
fluxes, or more precisely influxes to and effluxes from the canopy, for different N species,
namely ammonium, nitrate and DON (dissolved organic nitrogen). The terms efflux and
influx instead of uptake and leaching, respectively are more appropriate as we were not
able to determine dry deposition of nitrogen on canopy surfaces. In this approach
rainwater labelled with 15N the tracer is diluted by canopy leaching (efflux) and “consumed”
by canopy uptake processes (influx).
The 15N pool dilution experiment clearly demonstrated that high rates of gross N fluxes of
solutes occur across canopy components in tropical rainforests even though net fluxes are
minor or negligible (see Fig. 4, NH4+). Moreover, this also supports the dynamic model
therefore pointing to a greater importance of canopy exchange processes than evident
- 38 -
from net flux measurements alone. Canopies obviously are exchanging nutrients in a
bidirectional way (Wanek et al. 2003; Wanek and Pörtl 2005) and at higher rates than
previously anticipated which is analogy to foliar CO2- and H2O- exchanges that by far
exceed net CO2 assimilation and water losses (Heldt 1998; Helliker and Griffiths 2007).
Additionally the 15N-pool dilution approach revealed a net uptake of all examined N species,
ammonium, nitrate and DON (Fig. 3 and 4). This is in contrast to previous studies (Coxson
1991; Lovett and Lindberg 1993; Filoso et al. 1999; Wania et al. 2002) where throughfall
measurements in tropical forests showed that inorganic N, nitrate and ammonium, were
indeed taken up by the canopy, but DON leached from the canopy. Filoso et al. (1999)
suggested that large enrichments of DON in throughfall resulted from retention and
assimilation of deposited inorganic N on canopy surfaces, mainly by epiphytic algae, and
subsequent washing off or leaching as DON during later rain events. Further Coxson
(1991) reported that nutrients, mainly organic compounds, were released to a great extent
by epiphytic bryophytes. Due to the fact that bryophytes often provide a habitat for N2-
fixing bacteria and the occurrence of lichens with cyanobacterial symbionts (Bentley 1987)
release of N from atmospherically fixed nitrogen might provide another source of DON
within the canopy. On the basis of our data we cannot solve this discrepancy. DON such as
amino acids are readily taken up by plants as an N source which plants can immediately
utilize. Since in this experiment labelled DON was provided in the form of 15N-Glutamic
acid, canopy components such as cryptogamic epiphytes might utilize added DON in
rainwater. Although the N was supplied in a single pulse similar results were found for
three days of treatment with simulated rain events with ambient rainwater (Fig. 3) and for
a whole year period (Hofhansl, pers. commun.).
Even though net uptake of ammonium and nitrate did not differ (Table 2) gross influx of
ammonium (and DON) were about ten times higher than influx of nitrate. This may be due
to negatively charged sites in cell walls or cuticles e.g. carboxyl groups of pectins which
affect the passage of ions through the apoplast. The movement of ions in cell walls
(apoplast) is characterized by electrostatic interactions leading to preferred uptake of
cations (NH4+, positive charged amino acids) and repulsion of anions (NO3
-) (Clarkson
1993; Marschner 1995; Sattelmacher 2001).
At the first sight, the data with regard to net N fluxes from the second tracer experiment
disagree with results from the stem injection tracer experiment, since in the latter
experiment weak net leaching of ammonium, nitrate, and DON was found. However, it has
to be considered that the investigated canopies of the stem injection experiment only
consisted of foliage and small branches of woody plants and did not exhibit any visual
epiphytic or epiphyllous colonization whereas the branches of the second experiment were
richly covered with epiphytic bryophytes and vascular epiphytes and host tree leaves were
- 39 -
colonized by epiphylls. Thus, results of both experiments can not be directly compared
together but clearly demonstrate that epiphytic communities obviously play a major role in
solute fluxes in canopies and therefore in nutrient dynamics of tropical rain forests. This is
consistent with other studies (Nadkarni 1984; Chuyong et al. 2004) which observed higher
retention of N within the canopy with greater abundance of epiphytic bryophytes. Nadkarni
(1984) determined for a tropical cloud forest that the epiphyte biomass constituted less
than two percent of total forest dry mass but retained nutrients equivalent to 45% of that
in forest foliage and therefore might act as a “buffer” or an additional nutrient pool for
previous inaccessible sources and later release to other canopy components.
Net throughfall fluxes of DOC (dissolved organic carbon) were substantially higher in the
second experiment than in the first experiment. This might indicate that epiphytes also
have an important effect on DOC fluxes (Table 1 and Fig.3). According to other throughfall
studies in tropical wet forests (Coxson et al. 1992; Möller 2005; Schwendenmann 2005)
DOC generally shows the highest fluxes and concentrations in throughfall compared to
other solutes (Parker 1983) and implies leaching of organic compounds from both higher
and lower plant tissues such as sugars, polyols and organic acids (Filoso et al. 1999); pulse
release of these compounds were observed by Coxson et al. (1992) from desiccated
epiphytic bryophytes during initial rewetting events.
The 15N-pool dilution experiment also allowed evaluating net 15N of various canopy
components. Dry mass of the colonized branches mainly consisted of their leaves and
epiphytic bryophytes (mosses) (Fig. 5A). In fact, the main part of the epiphytic community
was made up of bryophytes which absorbed the greatest fraction of the applied 15N tracer
(16%, Fig. 5B) whereas host leaves did not retain more as other fractions (< 5%). Clark et
al. (1998) reported that epiphytic bryophytes and epiphytic assemblages accounted for
80% retention of atmospheric deposited inorganic N and Clark (2005) estimated that 33-
67% of wet deposited inorganic N was retained by these canopy components in a tropical
montane forest. 15N-uptake rates of canopy components were highest for epiphylls (mainly
comprised bryophytes), intermediate for other epiphytes (vascular epiphytes, lichens,
mosses), and lowest for tree leaves. Even bark, canopy soil and canopy litter showed net 15N retention. Wilson and Tiley (1998) determined in a Norway spruce forest that uptake of
N by bark was 2-10 times higher than in needles, due to more permeable barriers and
greater wettability. Bryophyte gametophytes are virtually “barrier-free”, due to the lack of
a cuticle and epicuticular waxes, which facilitates active and passive ion exchange
processes. Similar results were observed by Wanek and Pörtl (2005) in a 15N-pulse
labelling experiment where epiphyllous bryophytes quickly incorporated 15N. Nevertheless,
epiphytic and epiphyllous bryophytes exhibit rapid turnover times (Coxson 1991) and
therefore do not act as a long term storage for nutrients. Moreover, concerning their
poikilohydric character, desiccated bryophytes and lichens loose large quantities of
- 40 -
nutrients in the first minutes during rewetting (Gupta 1976; Coxson 1991). The lack of
epiphytes and epiphylls in canopies of the stem injection experiments might possible
explain why nitrogen was released in throughfall and emphasizes that different canopy
components may show very different characteristics with respect to solute exchange.
Wanek et al. (2003) reported that plant species may exhibit large differences in foliar
solute exchange with the environment as they showed extensive N exchange between
specific host plant species and epiphylls when plants were 15N labelled.
Net fluxes of other cations and anions of both experiments are consistently with other
throughfall studies (Forti and Neal 1992; Cavelier et al. 1997; Filoso et al. 1999; Chuyong
et al. 2004) with exception of chloride and sodium in the second experiment. Previously it
has been assumed that Na+ (and Cl-) behaves conservatively in tree canopies i.e. show not
net uptake or leaching and therefore can be used as a tracer for dry deposition (Hansen
1996; Draaijers et al. 1997). This has been the basis of the canopy budget model (Ulrich
1983) to estimate the relative contribution of dry deposition and canopy exchange to net
throughfall. Net uptake of Na+ on a short term and an annual basis (Fig. 3; and Hofhansl,
pers. commun) as well as significant leaching of Na+ in mangrove canopies cast significant
doubt on the applicability of the canopy budget model. However, the controls and
pathways of sodium and chloride remain obscure in this study.
In conclusion the study demonstrates that (1) in situ application of tracers allows to
investigate the fractional contribution of external sources (dry and wet deposition) and
internal sources (canopy exchange) of solutes in throughfall, (2) that canopy exchange is
controlled by bidirectional fluxes i.e. canopy efflux and canopy influx, and (3) epiphytic
communities play a major role in nutrient dynamics in canopies of tropical rain forests.
- 41 -
REFERENCES
BENGSTON, P., FALKENGREN-GRERUP, U., BENGTSSON, G. (2006). "Spatial
distributions of plants and gross N transformation rates in a forest soil." Journal of
Ecology 94: 754-764.
BENTLEY, B. L. (1987). "Nitrogen fixation by epiphylls in a tropical rainforest." Annals of
the Missouri Botanical Garden 74: 234-241.
BOWDEN, R. D., GEBALLE, G. T., et al. (1989). "Foliar uptake of 15N from simulated
cloud water by red spruce (Picea rubens) seedlings." Canadian Journal of Forest
Research 19(382-386).
BOYCE, R. L., FRIEDLAND, A. J., et al. (1996). "Direct canopy nitrogen uptake from
15N-labeled wet deposition by mature red spruce." Canadian Journal of Forest
Research 26: 1539-1547.
CAVELIER, J., JARAMILLO, M., et al. (1997). "Water balance and nutrient inputs in bulk
precipitation in tropical montane cloud forest in Panama." Journal of Hydrology
193(1-4): 83-96.
CHUYONG, G. B., NEWBERY, D. M., et al. (2004). "Rainfall input, throughfall and
stemflow of nutrients in a central African rain forest dominated by ectomycorrhizal
trees." Biogeochemistry 67(1): 73-91.
CLARK, K. L., NADKARNI, N. M., et al. (1998). "Atmospheric deposition and net
retention of ions by the canopy in a tropical montane forest, Monteverde, Costa
Rica." Journal of Tropical Ecology 14: 27-45.
CLARK, K. L., NADKARNI, N.M., GHOLZ, H.L. (2005). "Retention of Inorganic Nitrogen
by Epiphytic Bryophytes in a Tropical Montane Forest." Biotropica 37: 328-336.
CLARKSON, D. T. (1993). "Roots and the Delivery of Solutes to the Xylem." Philosophical
Transactions of the Royal Society of London Series B-Biological Sciences 341(1295):
5-17.
COXSON, D. S. (1991). "Nutrient Release from Epiphytic Bryophytes in Tropical Montane
Rain-Forest (Guadeloupe)." Canadian Journal of Botany-Revue Canadienne De
Botanique 69(10): 2122-2129.
COXSON, D. S., MCINTYRE, D. D., et al. (1992). "Pulse release of sugars and polyols
from canopy bryophytes in tropical montane rain forest (Guadeloupe, French West
Indies)." Biotropica 24(2a): 121-133.
DRAAIJERS, G. P. J. and ERISMAN, J. W. (1995). "A canopy budget model to assess
atmospheric deposition from throughfall measurements." Water, Air, and Soil
Pollution 85: 2253-2258.
DRAAIJERS, G. P. J., ERISMAN, J. W., et al. (1997). "The impact of canopy exchange
on differences observed between atmospheric deposition and throughfall fluxes."
Atmospheric Environment 31(3): 387-397.
- 42 -
FILOSO, S., WILLIAMS, M. R., et al. (1999). "Composition and deposition of throughfall
in a flooded forest archipelago (Negro River, Brazil)." Biogeochemistry 45(2): 169-
195.
FORTI, M. C. and NEAL, C. (1992). "Hydrochemical Cycles in Tropical Rain-Forests - an
Overview with Emphasis on Central Amazonia." Journal of Hydrology 134(1-4): 103-
115.
GARTEN, C. T., SCHWAB, A. B., et al. (1998). "Foliar retention of 15N tracers:
implications for net canopy exchange in low- and high-elevation forest ecosystems."
Forest Ecology and Management 103(2-3): 211-216.
GUPTA, R. K. (1976). "Physiology of Desiccation Resistance in Bryophytes - Intraspecific
Differences in Acrocladium-Cuspidatum (Hedw) Lindb." Indian Journal of
Experimental Biology 15(11): 1074-1075.
HAMBUCKERS, A. and REMACLE, J. (1993). "Relative Importance of Factors Controlling
the Leaching and Uptake of Inorganic-Ions in the Canopy of a Spruce Forest."
Biogeochemistry 23(2): 99-117.
HANSEN, K. (1996). "In-canopy throughfall measurements of ion fluxes in Norway