Unusually negative nitrogen isotopic compositions (?15N) of mangroves and lichens in an oligotrophic, microbially-influenced ecosystem

Extremes in δ 15 N values in mangrove tissues and lichens (range =+ 4 to − 22‰) were measured from a mangrove forest ecosystem located on Twin Cays, o ﬀ shore islands in Belize, Central America. The N isotopic compositions and concentrations of NH + 4 /NH 3 in porewater, rainwater, atmospheric ammonia, mangrove leaves, roots, stems, and 5 wood, and lichens, were examined to study the biogeochemical processes important for establishing these unusual N isotopic ratios. Porewater ammonium concentrations had little to no relationship to N isotopic fractionation in mangrove tissues. The δ 15 N of ﬁne and coarse roots was 9‰ more positive than leaf tissue from the same tree. When P was added to dwarfed mangrove trees without added N, δ 15 N increased within one 10 year to a δ 15 N closer to the δ 15 N of porewater ammonium ( δ 15 N =+ 4‰). Isotopically negative ammonia in the atmosphere ( δ 15 N = − 18‰) and in rainwater ( δ 15 N = − 9‰) were found on Twin Cays and may be sources of available N for isotopically depleted mangrove trees and lichens. In highly stressed, severely P limited trees, uptake of atmospheric N by Rhizophora mangle may be an important adaptive strategy. 15


Conclusions References
Tables Figures

Introduction
Mangrove trees, a wide range of species that can be characterized by their ability to grow in brackish or full salinity seawater along tropical and subtropical coasts, can be growth limited by N, or P or both (Feller, 1995;Feller et al., 1999).While it is straightforward to assume that N limitation will have some effect on the isotopic composition of plants (Evans, 2001), there are no clearly justifiable expectations about P limitation's effect on N isotopic compositions.P is taken up by plants through the roots, typically, assisted by the action of membrane bound phosphatases (Muchhal and Raghothama, 1999;Smith, 2001).Experiments on fertilized mangrove trees have shown that when a tree with the primary limitation of P is given this nutrient the tree responds rapidly and dramatically by adding shoots, roots, and new leaves (Feller, 1995).The δ 15 N of plants reflects the net effect of many processes including the δ 15 N of the source N, enzymatic fractionations within a plant, and plant-microbial interactions in soil (Dawson et al., 2002).The majority of terrestrial plants have δ 15 N near 0‰ in temperate zones, however, different species growing in the same environment have been found to vary by as much as 10‰ (Handley and Scrimgeour, 1997).Mangrove trees, in general, have N isotopic compositions that reflect the overall nutrient status of the ecosystem (Fry et al., 2000).For example, trees growing in regions associated with human populations had δ 15 N similar to that of the N released from sewage treatment plants.Accordingly, we have previously reported that Rhizophora mangle L. (Red mangrove) found on a frigate bird rookery only a few kilometers from the site of the present study had δ 15 N as positive as +17‰, which was identical to the N in the sediment and in the birds' droppings (Wooller et al., 2003b).
Plants can utilize N dissolved in soils or by absorption through their leaves (e.g., Garten et al., 1990;Leith et al., 2002) and volatile ammonia originating from animal colonies can influence the δ 15 N of plant leaves (Erskine et al., 1998).For example, negative δ 15 N (to −8‰) have been measured in grasses, C 3 plants, and mosses, which were collected growing downwind of major bird rookeries.Plants collected in the immediate vicinity of the colonies incorporated the enriched δ 15 N (up to +18‰) from the marine N at the base of the birds' food web.Either following a rainfall event or as the result of dry deposition, ammonia and nitrate can be taken up by plants and incorporated into their biomass.
The stable N and C isotopic compositions of mangrove trees fertilized with N or P have been studied previously (McKee et al., 2002).These authors, measuring differences in freshly collected leaves, found that R. mangle trees fertilized with P had positive δ 15 N values compared with unfertilized or N fertilized trees.The controls without fertilizer and the N fertilized trees had negative δ 15 N values to −8‰.In conditions where the concentration of ammonia or nitrate is in excess of what a plant needs for immediate growth, isotopic fractionations increase, then are displayed in the δ 15 N of plant tissues (Yonemaya et al., 1991;Fogel and Cifuentes, 1993)   Based on our previous work (Wooller et al., 2003a and b;Smallwood et al., 2003) and preliminary studies, a number of lines of evidence suggested that the explanation for depleted δ 15 N was incomplete.For example, Wooller et al. (2003b) noted that R. mangle leaves from unfertilized trees collected at random locations throughout the islands used by McKee and coworkers, had δ 15 N considerably lower than the values they had found, making the hypothesis of fractionation during uptake from the soil problematic even though porewater dissolved nitrogen values were much lower than those in fertilized plots.Subsequent work on the biochemical partitioning of the N isotopic signal in R. mangle leaves showed that the signal was not associated with any particular leaf component or biochemical fraction (Smallwood et al., 2003).Rather, differences between dwarf trees with low δ 15 N and tall or fringing trees with δ 15 N near 0‰ were similar in all chemical fractions.In this paper we further explored the causes of the wide variations in the stable N isotopic compositions in mangroves at Twin Cays, Belize, by sampling mangrove trees from two tree species, as well as organisms that rely primarily on atmospheric N sources (i.e., lichens), organisms that fix atmospheric N (e.g., microbial mats), and the isotopic compositions of potential soil, water, and atmospheric sources of this critical nutrient.

Study site and collections
Twin Cays is a highly oligotrophic, peat-based archipelago, located 12 km off the coast of Belize and approximately 3km inside the barrier reef.These islands are part of a Smithsonian research area, which has been the focus of long-term studies on mangroves, sponges, seagrasses, and coral ecosystems (Woodruffe, 1995; Ruetzler and Introduction

Conclusions References
Tables Figures Printer-friendly Version Interactive Discussion Feller, 1996;Wooller et al., 2004).Samples, including leaves, roots, bark, stems, and wood of mangrove trees (R. mangle and leaves only from Avicennia germinans (L.) Stearn.; black mangrove), microbial mats, and lichens were sampled during the period extending from October 2000 to April 2004 (Wooller et al., 2003a, b).Samples were dried at 50-70 • C at the Smithsonian Caribbean Coral Reef Ecosystem laboratory at Carrie Bow Cay, approximately 3 km from Twin Cays.

Fertilization experiments
For the N and P fertilization studies, we sampled the three fertilization sites established in January, 1998, extending from the fringe into the dwarf zone (Feller et al., 2003).I.C.Feller and coworkers of the Smithsonian Institution maintained these fertilization plots.In addition, 12 dwarf red mangrove trees were chosen for a one-time experiment.Three trees served as controls; three trees received 150 g of P 2 O 5 (0:20:0) ,which were buried within 1cm of a major prop root; three trees were fertilized with 150 g of P 2 O 5 (0:20:0), 1 m away from the nearest prop root.This fertilizer aliquot was located so that no other mangrove tree was within 2 m.Leaves were sampled periodically over the next several years for isotopic and elemental compositions; trees were also assessed for internodal length, production of new prop roots, and overall size.

Bulk C and N stable isotope analyses
For leaf bulk stable C and N isotope analyses, a small aliquot of dried leaf tissue was sampled, taking care that no veins, invertebrate or fungal damage, or discolored portions of the leaf were used.Lichens were scraped off of bark with spatulas or razor blades taking care not to include bark tissue in the sample.An aliquot (∼400 to 1000 µg) of each sample was analyzed using continuous-flow, stable isotope ratio mass spectrometry (Finnigan MAT, Delta plus XL) (as described in Wooller et al., 2003b

Ammonium concentration and isotopic measurements
Porewater ammonium concentrations were determined by collecting 10-50 ml at 10 cm depth using a sipper system.Water samples were filtered at the laboratory on Carrie Bow Cay with glass fiber filters (GF/F).Filtered water samples were stored in the refrigerator and analyzed typically within 2 days following collection.For samples that contained H 2 S, 0.2 ml of HCl was added to the porewater prior to analysis following the method of Solarzano (1969).
To measure δ 15 N of ammonium, we collected 120 ml of porewater from 5-10 cm depth using sippers, filtered and analyzed as above.Samples were collected from multiple locations around Twin Cays.Filtered porewater (110 ml) was decanted into a specimen cup with a Teflon sandwich containing filter paper dosed with 5-10 ml of H 2 SO 4 -KH 2 SO 4 solution following a modification of the method of Stark and Hart (1996).

Atmospheric ammonia emissions and isotopic measurements
For measurements of ammonia emissions into air, we deployed ammonium sensitive badges (K&M Environmental, Virginia, USA) that absorbed ammonia gas through a Teflon membrane, producing a color change by reacting with an ammonia sensitive chemical indicator (see Kring et al., 1981).The average ammonia emission over the period of exposure is estimated by dividing the dose by the time of exposure (i.e. the data reported by the badges are in units of ppm NH 3 -hr).Badges were deployed in the field by tying them onto mangrove prop roots or stems at a height of 20-50 cm above the mean high tide.Readings were taken over time up to 24 h.To collect ammonia in the field for isotopic analyses, we deployed Teflon sandwiches (see above) contained in mesh bags, which were tied to mangrove branches or PVC pipes around the islands during February 2003 and March 2004.The tags were left out for a period of 6 days, which is considered adequate for absorbing atmospheric ammonia gas (Roadman et al., 2003).Introduction

Conclusions References
Tables Figures Printer-friendly Version Interactive Discussion

Laboratory flux experiment
In March 2004, we conducted flux studies at the Carrie Bow Marine Laboratory using peat or mat cores (diameter=5 cm) collected from the field the previous day.Experiments were started before sunrise by hanging a K&M badge (see above) inside of the core barrel and incubated the cores in full daylight at ambient air temperatures ( 26to 29 • C).Ammonia readings were taken every hour for 6 hours, or until the ammonia detecting badges reached saturation (≪300 ppm NH 3 ).

P concentrations
Porewaters for P analysis were collected, filtered and frozen for subsequent measurement in Los Angeles on a Latchatt Automatic Analyzer at USC.Total P was determined on leaves by combusting approximately 100 mg of leaf material at 500 • C for 2 h.The resulting ash was weighed, digested using methods modified from Jensen and Thamdrup (1998), and then analyzed for P with spectrophotometric methods (Presley, 1971).

Stable isotopic signature of mangroves from Twin Cays, Belize
The dominant tree species on Twin Cays, Belize, R. mangle (red mangrove), had an unusually large range in bulk leaf δ 15  The greatest variations of δ 15 N were found in the leaves of R. mangle located in the islands' interior.Very negative δ 15 N (<−8‰) were measured only in dwarf (i.e.<1 m tall) or interior tall trees (Table 1).Many of these dwarf trees had very short internodal lengths (<0.2 cm) reflecting their slow growth.Isotopic heterogeneity was extreme over very short distances (e.g., 2 m or less).For example, dwarf trees found within 1 m of interior tall trees often differed from each other in terms of δ 15 N by up to 14‰.
The δ 15 N of fringe R. mangle was less variable than interior trees (Table 1).The δ 13 C of fringe trees were more positive on the exterior, wave-beaten portions of the island than in the inner channels (p<0.001).
The results of isotopic analysis of a single dwarf tree, sacrificed in its entirety, from the Batfish Pond site are summarized in Table 3.Although it was initially chosen at random, it represented one of the many examples of the extremely low δ 15 N trees.The δ 15 N of above ground tissues had little variation between flowers, buds, leaves, stems, wood, and prop roots.Below ground tissues varied in δ 15 N from −3.1‰ to −15.6‰, most noticeably within the roots which are actively involved in nutrient and water absorption.Similar results were obtained in less complete sampling of unfertilized dwarf trees from other Twin Cays sites.Striking in all cases was the absence of variation or gradients in δ 15 N through the shoot, from the substrate surface to the most distal leaves, or with leaf developmental stage.Similarly noteworthy were the higher and variable values of δ 15 N in the active roots.

Stable isotopic composition of microbial mats and lichens
The isotopic composition of bacterial and algal mats distributed on the islands had δ 15 N with an average of −0.9±1.3‰(range=+2.3to −3.7; n=90) (Fig. 2).These δ 15 N values indicate a source of N from N fixation (e.g., Macko et al., 1986).Lichens had a δ 15 N range from 0.4‰ in the fringe zone to −21‰ in the floc and dwarf zones (Fig. 3).
The δ 15 N of the lichens was not related to the δ 15 N of the bark that the lichens were found growing on (r 2 =0.167) or leaves growing on nearby branches.Patches of lichens both the bark and a corresponding leaf that was collected at the same tree height from the sediment surface.

Fertilization experiments
In the long-term fertilization plots (Feller et al., 2003) at the Dock, Boa Flats, and Lair sites, trees fertilized with P had leaf δ 15 N that were around 0‰ with no significant difference across the tree height gradient.Dwarf trees fertilized with the addition of N as urea (δ 15 N=0‰), had δ 15 N values as low as −12.4‰, whereas those from fringe and transition trees likewise fertilized, were more positive (Tables 2, 4).At the Dock site, some of the N fertilized trees as well as control trees measured in 2003 had more negative N isotopic compositions than the same trees measured in 1998 (McKee et al., 2002).The δ 15 N in P fertilized tree leaves was as positive as 2.6‰ at this time, similar to δ 15 N in unfertilized fringe trees.
In 2002 a one-time, P fertilization experiment was started to test how quickly, and to what extent, a single dose of P can affect the growth and functioning of dwarf R. mangle trees.Trees that received P input directly next to a major prop root displayed stimulated growth within 7-8 months, as evidenced by increased internodal distances (from 0.1cm to >5cm) and more positive δ 15 N (Fig. 4).Trees that were fertilized with P one meter away from major prop roots experienced a lag phase relative to those trees fertilized proximally.The δ 15 N and δ 13 C and the growth of the control trees, approximately 25m away from the P fertilized areas, remained constant during the experimental period.Ammonia emissions in situ (February 2003 andMarch 2004) were measured coincident with porewater concentrations.They were highest over mats and floc (Table 5; Fig. 5).In February 2003, we sampled the δ 15 N of atmospheric ammonia, ammonium from underlying porewaters, and rainwater (Table 6).Both atmospheric ammonia and the ammonia/ammonium in rainwater on Twin Cays had very negative isotopic compositions relative to those measured in porewaters and in rainwater collected on Carrie Bow Cay.The isotopic fractionation between the atmospheric ammonia (Mean δ 15 N=−18.5‰) and the ammonium in porewaters (Mean δ 15 N=5.3‰) is 23.8‰, which fits within the range of isotopic fractionations (19 to 30‰) between these two N species that have been measured during chemical isotope fractionation experiments in the laboratory (Thode et al., 1945;Hermes et al., 1985).
Flux experiments with small cores incubated on Carrie Bow Cay showed that ammonia fluxes related positively to porewater concentrations (r 2 =0.7). 4 P and N concentrations and ratios in red mangrove leaves and underlying sediments P concentrations in freshly collected R. mangle leaves were extremely low all over Twin Cays (0.06±0.02%Total P (TP); n=71).Red mangrove leaves sampled on the East Island remote from direct coastal access and completely removed from fertilization plots, had 0.08±0.04%TP with ranges from 0.03 to 0.20 %TP (n=35).These %TP values were elevated slightly from trees receiving chronic, episodic P fertilization (0.07±0.03%TP; n=27) (p=0.04) in the three fertilization experiments maintained on the islands (Feller et al., 2002).Tissue levels, thus, were constant in fertilized trees, but total tree levels were much higher as many more leaves, branches, and stems were produced.Trees receiving a one-time input of P had significantly more total P (0.10±0.05%TP) in growing leaves 2 years after the initial fertilization (p=0.05)than control trees (0.07±0.02%TP).

Discussion
Nitrogen isotopic compositions of mangrove tissues were not simply related to the inorganic N concentrations in sediments.Although it has been postulated that higher sediment N concentrations would result in increased N isotope fraction during uptake and biosynthesis in roots, based on the ammonium concentrations in porewaters and the coexisting N isotopic compositions in mangrove leaves, we find minimal correlative evidence to support this statement (R 2 =0.31).For example, dwarf trees fertilized with urea had higher porewater ammonium concentrations (e.g., 200 µM) but with δ 15 N values down to −10‰, compared to unfertilized dwarf trees growing in sediment with 20-30 µM ammonium with δ 15 N as low as −18‰.
We found that trees fertilized with P and some trees growing nearby with equivalent sedimentary and leaf total P concentrations had the most positive δ The relative amount of P then seems to be the most important factor in determining δ 15 N values.However, the relationship between total P or N:P in leaves or sediments and δ 15 N of leaves was not, a linear or exponential function (Fig. 6).Clarkson et al. (2005) recently measured a correlation between % total P in foliar tissue and δ 15 N in plants growing in peat bogs.They concluded that P limitation reduced mycorrhizal colonization of roots, which in turn potentially resulted in increased N isotope fractionation.Clarkson et al. (2005) Hobbie et al. (2000Hobbie et al. ( , 2005) ) have shown that a substantial N isotopic fractionation occurs during early colonization by ectomycorrhizal (ECM) and ericoid mycorrhizal (ERM) fungi associations with δ 15 N in leaves of colonizing species averaging -9 ± 2‰.Mature plants infected with AM fungi had δ 15 N has low as -4‰; our leaves with δ 15 N less than −10‰ were collected from dwarf mangrove trees that were well established.Mangrove ecosystems at Twin Cays have not shown evidence of mycorrhizal associations with mangrove trees (Lovelock, personal communication), although arbuscular mycorrhizal (AM)-type mycorrhizal associations have been found in mangroves colonizing the Ganges River estuary (Sengupta and Chaudhuri, 2002).It is unlikely that AM fungal associations could explain the full extent of the δ 15 N variation we determined at Twin Cays.
The intimate relationship between P availability and N isotopic fractionation was shown clearly and dramatically in the results of fertilization studies in as little as 3 months (Fig. 4).After 5 months, the δ 15 N increased from −14‰ to −9‰ in leaves that had P applied directly at the roots, as opposed to 1m away or sprayed on the leaves.After 8 months, all P fertilized trees, regardless of the application method, Although the potential sources for N for mangroves could include dissolved nitrate or amino acids in the sediment, the negative δ 15 N is likely explained by the incorporation of an atmospheric NH 3 pool.Nitrate levels in unamended surface sediments from Twin Cays were rarely above 25% of the total dissolved N pool, and there were no indications that any sediment pool had a δ 15 N different from 0‰ (data not shown).Amino acids in porewaters are another potential source of N for mangroves, although their uptake by mangroves has not been specifically documented.In a review by Lipson and Nasholm (2001), they state that in ecosystems where microbial activity and biomass have large seasonal cycles, organic N uptake might be a significant source of N. Again, however, the total sediment δ 15 N, together with the decreased 15 N depletion in the roots compared to the leaves suggests amino acid uptake was not a significant source.Ammonia in the atmosphere was always depleted in 15 N, with isotopic compositions averaging -18±4‰ (n=20).Ammonium in rainwater (δ 15 N=−9±3‰; n=4) was also depleted in 15 N. Thus, uptake of N from an atmospheric or rainwater pool is a potential explanation of the δ 15 N of the most negative dwarf trees.This is further supported by the results of lichen analyses.The δ 15 N of lichens collected from trees in the dwarf, transition, floc, and fringe zones of the island also have very negative values (as low as δ 15 N of −22‰).Lichens, and other epiphytes, do not have roots, thus must use atmospheric sources of N for their nutrient requirements; therefore the possibility that an isotopically depleted N source is available to them for growth is strong (Hietz et al., 2002).We propose that N isotope fractionation may be a passive process related to leaf proximity to volatilized ammonia and is proportional to the relative amount of

BGD Introduction Conclusions References
Tables Figures Full Screen / Esc

Printer-friendly Version
Interactive Discussion uptake of N by the roots (Fig. 7).Foliar uptake of ammonia might be an important, critical source of N for the Twin Cays mangrove ecosystem, as has been shown with terrestrial vegetation particularly in polluted areas (e.g.Krupa, 2003).This foliar uptake may be especially important in trees which are P limited, as it has been shown that P limited mangrove trees put less energy into below ground biomass (e.g., McKee et al., 2007).

Conclusions
We conclude that P-limited dwarf mangrove trees, growing adjacent to ammonia sources, in addition to lichens on trees in floc and dwarf zones, obtain a portion of their N from atmospheric sources (i.e.ammonia) with isotopically distinct isotopic compositions from porewater N. The very negative δ 15 N of mangroves and lichens on the oligotrophic islands of Twin Cays are key in estimating the primary and secondary nutrient limitations for this ecosystem.Twin Cays, located away from terrestrial runoff and with limited outside atmospheric influences, is an area where dwarf mangrove trees struggle to survive.We propose that their success depends in part on their taking advantage of the microbial community comprised of N-fixing cyanobacteria and other photosynthetic microbes.A small, but significant portion of the ammonia fixed by microbes is released by physical processes into the atmosphere where it is available for uptake by leaves on trees growing in high salinity, low P, anoxic sedimentary porewaters.This strategy must also depend on leaf proximity to volatilized ammonia sources, such as exists at Twin Cays, and the resulting δ 15 N reflects the relative amount of uptake of N by leaves and roots.Thus, foliar uptake of ammonia is a critical source of N for the Twin Cays mangrove ecosystem.Finally, we predict that in this and other highly oligotrophic ecosystems, δ 15 N can be powerful indicators of the integrated P and N cycling.
From microbial mats and floc samples, rates of 2.0 to 3.0 µmole NH 3 /m 2 /hr were measured in comparison to fringe cores, which had an average flux of 0.7 µmole NH 3 /m 2 /hr.

FringeFig. 1 .
Fig. 1.The δ 15 N and δ 13 C of Rhizophora mangle leaves on Twin Cays, Belize.All data from these trees were collected from trees not included in specially fertilized experimental plots.(a) Interior Tall and Dwarf Trees.(b) Fringe Trees.
evidence of responding by having longer internodes, thinner leaves, more leaves, new prop roots, and reproductive tissues.Control trees were unchanged.At the same time, δ15N in leaves increased in all P treatments: direct root application, δ 15 N=−2‰ and 1 m distant root application δ 15 N=−4‰.In conclusion, P availability can explain the continuum of δ 15 N values from negative to positive values, but not the showed

Table 2 .
Statistical analysis of Rhizophora mangle fresh leaf isotopic compositions as a function of stand structures.

Table 3 .
Isotopic compositions of tissues from a single dwarf tree from the Batfish Pond region."Functioning prop root" indicates that it has penetrated the substrate and produced one or more classes of finer roots."Smaller roots" were in contact with the substrate, ranging in diameter from <0.5 mm to >8 mm.

Table 4 .
N isotopic composition of fresh leaves from fertilized R. mangle experimental plots on Twin Cays, Belize.Samples were collected and measured in 2003.

Table 5 .
Ammonia emissions from Twin Cays.Floc refers to unconsolidated microbial growth floating on surface peat sediments.Floc often occurred in thick (>1 m) banks forced to the edges of interior ponds by wind and waves.Microbial mats are consolidated sedimentary features with distinct layering (seeJoye and Lee, 2004).

Table 6 .
Isotopic compositions of ammonium plus ammonia from porewaters, rain, and atmosphere from Twin and Carrie Bow Cays, Belize February 2003.*The δ 15 N standard was 0.1‰.