Leaf Gas Exchange and Nutrient Use Efficiency Help Explain the Distribution of Two Neotropical Mangroves under Contrasting Flooding and Salinity

Rhizophora mangle and Laguncularia racemosa cooccur along many intertidal floodplains in the Neotropics. Their patterns of dominance shift along various gradients, coincident with salinity, soil fertility, and tidal flooding.We used leaf gas exchangemetrics to investigate the strategies of these two species in mixed culture to simulate competition under different salinity concentrations and hydroperiods. Semidiurnal tidal and permanent flooding hydroperiods at two constant salinity regimes (10 g L and 40 g L) were simulated over 10 months. Assimilation (A), stomatal conductance (g


Introduction
Mangroves are forested wetland habitats characterized by individual tree species that can segregate along rather diffuse environmental gradients.Salinity and hydroperiod (i.e., frequency, depth, and duration of flooding) are considered two of the more important factors that control the local distribution of mangrove species [1][2][3][4][5], and both factors have the potential to interact strongly along spatial and temporal scales in many floodplains [6][7][8][9].Indeed, interactions between salinity and hydroperiod have been implicated in the regulation of survival, growth, physiological proficiency, and zonation of individual seedlings, saplings, and trees that colonize mangrove wetlands [10,11].Through repetitive experimentation, we are progressively gaining a better understanding of the interactive nature of these environmental drivers on seedling establishment and competitive outcomes within floodplain settings globally [12].
Rhizophora mangle L. and Laguncularia racemosa (L.) C. F. Gaertn.frequently coexist and are common mangrove species along coastal reaches of riverine floodplains in the Neotropics [13][14][15].Rhizophora mangle can occupy a broad range of tidal flooding, from highly abundant in lower intertidal zones with high annual flood durations (up 50-60%) to less abundant in upper intertidal zones with lower annual flood durations [4,16,17].L. racemosa is often found International Journal of Forestry Research coexisting with R. mangle, but the occurrence of L. racemosa can favor landward extremes where salinity might at times be higher than in lower intertidal areas [14,18].In addition to salinity [19,20] and hydroperiod [15,[21][22][23], the distribution of L. racemosa and R. mangle is also correlated with soil resources such as nitrogen (N) and phosphorus (P) [24][25][26].All site factors interact considerably along such riverine flood and tidal gradients, confusing our understanding of physiological preference in Neotropical coastal floodplains.
In this study, we tested whether leaf gas exchange responses in seedlings of R. mangle and L. racemosa differed under varying conditions of salinity and hydroperiod to understand the role of these drivers in mediating the competitive interactions among these two mangrove species at an early growth stage.Proficiency in assimilation () of carbon as CO 2 during photosynthesis can set the upper limit of individual tree growth and ecosystem productivity [27].Species-specific differences in  can also be suggestive of among-species competitive relationships [28][29][30].Lovelock and Feller [18], for example, attributed a switch in dominance between Avicennia germinans (L.) L. and L. racemosa along salinity and nutrient gradients to leaf-level gas exchange strategies and the trade-off between photosynthetic water use efficiency (PWUE) and photosynthetic nitrogen use efficiency (PNUE).We opted for a similar investigation here but include the competitive interactions between L. racemosa and a different species (R. mangle) for comparison.

Plant Material.
Propagules of R. mangle and L. racemosa were collected from Terra Ceia Island near Bradenton, Fl USA (27 ∘ N, 82 ∘ W).Propagules were transported to Lafayette, LA, USA (30 ∘ N, 92 ∘ W), sown directly into 0.4 L plastic cups with sand, irrigated with fresh water to the soil surface, and, after seedling emergence, were replanted into 27 L buckets.Soil in the buckets consisted of an equal mix of river sand and commercial potting mix (Sta-Green, Spectrum Brands, Atlanta, GA, USA).One-half of the seedlings were subjected to a 12-week acclimation period at 10 g L −1 salinity with 10 cm of surface flooding.The remaining seedlings were acclimated identically until the last 20 d when salinity was gradually adjusted to a final concentration of 40 g L −1 .Water levels were adjusted from 10 to 20 cm by an increment of 1.5 cm d −1 .Thereafter, the entire volume of experimental water was replaced every 2.5 months.Initial mean heights of seedlings were 35.8 ± 0.4 (s.e.) cm for R. mangle and 40.1 ± 0.7 (s.e.) cm for L. racemosa at the point that treatments were imposed and competition began.
Average irradiance in the greenhouse was approximately 41% of ambient sunlight.Maximum instantaneous photosynthetically active radiation (PAR) in the greenhouse was approximately 800 mol photons m −2 s −1 and generally did not appear limiting to leaf-level photosynthesis [31].Greenhouse temperatures were maintained above 8 ∘ C during winter months and below 43 ∘ C during summer months (mean of 22 ∘ C during experiment).Relative humidity ranged from 45 to 90%.

Experimental Design.
The experiment was maintained for approximately 10 months from August to May.Plants used for gas exchange measurements were selected from a larger experimental setup in which seedlings of both species were competing in paired seedling arrays (1 L. racemosa; 1 R. mangle) within each 27 L bucket [32].Buckets were located on tables, which produced flooded hydroperiods (permanently) and tidal hydroperiods by circulating water through reservoirs below.Each table contained six buckets: three were subjected to flooded hydroperiods and three were subjected to tidal hydroperiods.Tables were assigned in pairs to salinity of either 10 g L −1 or 40 g L −1 for a total of six tables per salinity treatment.Salinity concentrations were selected as approximate average (10 g L −1 ) and approximate maximum (40 g L −1 ) concentrations recorded from floodplain mangroves along the lower Shark River, Fl, USA [22].The entire experimental design was replicated six times on 12 independently maintained tables for a total of 72 individual seedlings per species.Maximum water level for each hydroperiod treatment was 20 cm.Tidal hydroperiods were simulated as two floods per day (0900; 1600), with each tide flooding seedlings for approximately 6 h to simulate lower intertidal conditions along the Shark River ( [23] site SRS-6).Soil oxidation-reduction (redox) potentials ranged from −48 to 149 mv for tidal hydroperiods and from −112 to −17 mv for flooded hydroperiods (mixed culture experiment [32]).

Photosynthetic Measurements.
Photosynthetic gas exchange characteristics were determined for 48 seedlings of L. racemosa and R. mangle during two sampling months (September, November).Subsamples of two experimental buckets per hydroperiod treatment (permanently flooded versus tidally flooded) were chosen randomly, as was the daily measurement order, which yielded 24 sample replicates (2 salinities × 2 hydroperiods × 6 experimental replicates) per species per month.From each selected seedling, one leaf was systematically selected from either the highest node in a lag phase of growth as determined by a fully developed apical sheath atop the node or, if damaged or obscured, from the node below.Of the two leaves per nodal pair, we selected the best looking leaf.
To document bias associated with measuring seedlings under either flooded or drained states during respective hydroperiod treatments (sensu [31]), gas exchange measurements in tidal hydroperiods were made under both states.Parallel measurements were conducted on seedlings grown in flooded hydroperiods for comparison (under both salinities).There were no differences in the leaf gas exchange responses evaluated in this study, while tidal treatments were flooded versus drained ( = 0.05), so the distinction was collapsed statistically for remaining analyses (see below).All measurements were made in random but variable order by treatment between 0900 and 1330 to account for response variation over the 4.5 h period.
Each leaf was placed inside the cuvette of a LI-6400 Portable Photosynthesis System (Li-COR, Lincoln, NE, USA).Conditions inside the cuvette were selected as follows: For comparison, sapling leaf temperatures for R. mangle along a south Florida estuary in October and December of a previous year ranged from 26.6 to 33.5 ∘ C (unpublished data from [31]).Depending on individual leaf response, each sample was allowed 2 to 8 min to equilibrate to a steady state for  (assimilation rate) and   (stomatal conductance).We included the most relevant variables in this study: ,   ,   (transpiration), and intercellular CO 2 concentration (  ).PWUE (instantaneous) was calculated as the quotient of /  .We also measured PNUE by determining leaf nitrogen concentration with an elemental analyzer (Model NC2500, Carlo-Erba Elemental Analyzer, Lakewood, NJ, USA).Leaves were harvested at the end of the experiment, dried to a constant weight at 60 ∘ C, and ground prior to nitrogen analysis.PNUE was calculated as the quotient of  and leaf nitrogen concentration, was destructive, and, hence, determined only for November for 6 leaves per treatment combination (i.e., overall sample size of 24 leaves per species).

Statistical Analysis.
Data were analyzed with an analysis of variance (ANOVA) as a split-split-plot design with salinity designated as the whole plot, and hydroperiod and species designated as the sub-and sub-subplots, respectively.Month was also treated as a subplot; repeated measures could not be used because the sample unit (i.e., individual plant) changed between months, even though the experimental unit (i.e., table) remained constant.Two separate analyses were conducted on the overall dataset, both included species and hydroperiod as fixed effects and salinity as a random effect.,   ,   , and PWUE were analyzed as a general linear model using ANOVA with month (September versus November) specified as a block.PNUE was determined only in November and was analyzed similarly to the other variables but with month collapsed using mixed model procedures.
Prior to all analyses, data were inspected for normality in the distribution of their residual errors and transformed when appropriate.All graphs report untransformed means.Analyses were performed by using either Proc GLM or Proc Mixed of SAS version 9.1 (SAS Institute Inc., Cary, NC, USA).

Assimilation and Intercellular CO 2 Concentration
. At the leaf level,  did not differ between R. mangle and L. racemosa (Table 1).Salinity and month, but not hydroperiod, affected assimilation rate, even though there was a significant interaction among species, salinity, and hydroperiod (Table 1).Higher values of  were generally obtained at lower salinity, but these effects were not attributable to specific treatment combinations after Bonferroni adjustment (Figure 1). was higher in September (Table 1) and decreased by 34.2% for L. racemosa and 33.7% for R. mangle from September to November, perhaps reflecting photoperiod, temperature, and/or relative humidity sensitivity.
Species did show significant differences in   , but did not display the significant interactions among species, salinity, and hydroperiod as for  (Table 1).Higher   was registered by L. racemosa at 10 g L −1 salinity (Figure 2).  was not significantly affected by hydroperiod, but   was slightly depressed with experimental duration which ultimately affected .Seasonal responses by the two species were also similar, that is, consistently lower   in November as ambient temperature during sampling dropped by 3.6 ∘ C and PAR increased by 146 mol photons m −2 s −1 .
There was a significant correlation between  and   for L. racemosa ( < 0.001; Figure 3  for R. mangle suggest that R. mangle tends to have higher  at a given   under all treatment combinations tested.

Stomatal Conductance, PWUE, and PNUE.
Mean   was significantly greater in L. racemosa than in R. mangle across all combinations (Table 1).As was the case for , these effects were not clear among specific treatment combinations after Bonferroni adjustment (Figure 1).Again, salinity and month significantly affected stomatal response, whereas hydroperiod did not (Figure 1), even though the two hydroperiods simulated had different effects on   at contrasting salinities for both species as indicated by significant salinity × hydroperiod interactions (Table 1).It is also important to stress the strong monthly response on   .Higher overall values were recorded in September (0.115-0.279 mol m −2 s −1 ) than in November (0.063-0.138 mol m −2 s −1 ) for both species.
There was a decrease in   in both species between months, with differences for L. racemosa and R. mangle remaining comparable (0.13 versus 0.10 mol m −2 s −1 , resp., or a 58% versus 56% decrease).PWUE was significantly higher in R. mangle than in L. racemosa at low salinity (Figure 2).Not only was there a difference between species and sensitivity to salinity and month, but also there was no effect of hydroperiod on PWUE (Table 1).L. racemosa had significantly higher PWUE at higher salinity (Figure 2), where PWUE increased from 2.64 ± 0.64 (s.e.) mol CO 2 (mmol H 2 O) −1 at salinities of 10 g L −1 to 3.27 ± 0.77 (s.e.) mol CO 2 (mmol H 2 O) −1 at salinities of 40 g L −1 .
Finally, L. racemosa exhibited significantly higher overall PNUE than R. mangle (Table 1), with differences becoming most disparate under tidal treatments at 40 g L −1 (Figure 4).Rhizophora mangle (R) in response to 10 g/L (10S) or 40 g/L (40S) salinity.Data are displayed according to the highest order significant interaction (Table 1).Values are means ± 1SE.Effects annotated with the same letter indicate lack of significance ( > 0.05) after Bonferroni adjustment ( = 0.05/12) for a particular graph.Both species generally had higher PNUE at low salinity (Table 1) by tending to concentrate more foliar N at the higher salinity concentration; foliar N concentrations displayed inconsistent responses between tidally and permanently flooded hydroperiods (Figure 4).

Interspecific and Seasonal Differences in Leaf Gas Exchange.
Average  for R. mangle in this study was comparable to the 7.5 to 14.2 mol CO 2 m −2 s −1 reported by others under similar light levels and salinity concentrations [30,31,[33][34][35].Accordingly,  for L. racemosa, which averaged 9.6 ± 0.29 (s.e.) mol CO 2 m −2 s −1 , was also comparable to field results of Lovelock and Feller [18], where  averaged 7.7 ± 0.6 mol CO 2 m −2 s −1 for trees fertilized with P and 9.5 ± 0.4 mol CO 2 m −2 s −1 for trees fertilized with N. On the other hand,   for R. mangle in our study (0.06-0.24 mol m −2 s −1 ) can only be compared to lower ranges reported by others [30,[36][37][38] and highlights the importance of intrinsic water conservation strategies in R. mangle as salinity is increased under different hydroperiods in experimental culture and presumably field settings.
appeared to be only loosely controlled by   in L. racemosa during gas exchange measurements, thus reflecting a short-term buildup of intercellular CO 2 as sustained high light levels during measurements opened stomata acutely to CO 2 diffusion.  ranged from 138 to 285 mol CO 2 (mol air) −1 in L. racemosa and from 107 to 310 in R. mangle across both months;   for healthy mangrove leaves are typically around 170 mol CO 2 (mol air) −1 [39].Yet, it is often noted that stomatal conductance has a strong effect on photosynthesis [40,41], and   and   appeared to be synchronic in our study; dedicated studies of specific /  y = (0.0047)e 0.0789x ; r 2 = 0.8961 responses would be more conclusive (Figure 3).Apparently, however, the occurrence of higher   does not guarantee higher carbon assimilation if the photosynthetic capacity of the species or imposed environment is low [42].L. racemosa certainly grew faster and to a larger size in this mixed-culture experiment than R. mangle [32].Interspecific differences in the sensitivity of   with relative humidity can also be an important determinant of stomatal behavior.  increased exponentially with an increase in the relative humidity inside of the leaf cuvette across both species in our study (Figure 5).Ambient temperature at the leaf surface of seedlings in September averaged 3.6 ∘ C higher than in November and likely affected the seasonal drop in leaf gas exchange.Mangrove leaf temperatures can exert an important influence on photosynthesis independent of stomatal activity [39,42].At physiologically acceptable low temperatures, an increase in temperature leads to spikes in , but any increase above temperature thresholds causes a strong decrease in  [43].This threshold for mangroves is about 35 ∘ C [10,39]; leaf temperatures in this study were measured to a maximum level of 37.2 ∘ C (mean, 32.8 ∘ C) during September but only to a maximum level of 31.5 ∘ C (mean, 29.2 ∘ C) during November.

Leaf Gas Exchange and Salinity.
Salinity has been identified repeatedly as constraining mangrove plant community growth and development at different concentrations depending on species [44,45].Mangroves have evolved conservative water use strategies at both the leaf level [10] and individual tree level [46,47], which are especially evident at higher salinity concentrations (e.g., [48]).Low stomatal conductance is characteristic of this strategy as freshwater loss through stomata is balanced against limited CO 2 gain [49].Species more tolerant to higher salinity are typically more efficient in water use [50].When growing in high salinity habitats, for example, the mangrove A. germinans exhibits lower   and higher PWUE than does L. racemosa, which is typically regarded as less salt tolerant than A. germinans [18].Certainly, growth increment was more affected by salinity in L. racemosa than in R. mangle in mixed culture here [32].Also, most species in the genus Rhizophora tend to increase PWUE with increasing salinity [29]; this increase has a metabolic cost to mangroves because of the unavoidable reduction in CO 2 assimilation [10,49,50].In our study,   was affected by salinity differently under contrasting hydroperiods (Table 1), while PWUE was not.
PWUE increased at higher salinities in our study, but PNUE dropped appreciably as salinity increased in concentration by 30 g L −1 (Figure 4).Recent studies have suggested a similar lowering of PNUE with salinity [18,29,51,52] and have indicated that individual species possess different limits to absolute rates of either PWUE or PNUE depending upon growth condition [18,31].Actual foliar N concentrations, as a component of PNUE, reflected greater foliar loading of N at higher salinities in both species as a priority adjustment in lieu of enhanced  in our study (Figure 4).Indeed, the relative proportion of soil N and P concentrations have been found to differ along natural gradients as well [22,53], and the relative importance values for R. mangle versus L. racemosa dominance also shift along the same gradient.For example, at locations of 1.9 and 4.1 km upstream from the mouth of the Shark River along the floodplain (south Florida), R. mangle and L. racemosa share dominance.But farther up river, R. mangle dominates at 9.9 and 18.2 km as soil N/P ratios increase with a decrease in soil total P relative to soil total N [22].Yet, N availability seems most important.Near the mouth of the river in higher salinity zones and also higher concentrations of total P, N is 2-4 times more available than in the lower salinity zones of the floodplain (18.2 km in land).This increase in inorganic nitrogen was attributable to higher ammonification rates measured near the mouth of the river and is correlated with phosphorus fertility of mangrove soils [22].
Our research suggests that the differences in PNUE for the two species may also contribute to patterns of community dominance along a coastal floodplain gradient as salinity changes.Where salinity is higher, greater leaf-level PNUE for L. racemosa might facilitate competitive dominance over R. mangle until exchangeable N increases.Enhanced competition between the two species might be especially important in zones of higher salinity and higher relative P, whereby ammonification is facilitated (cf., [22]), and more inorganic N becomes available for storage within leaf tissue.Lovelock and Feller [18] found similar results for L. racemosa in Florida (low PWUE; high PNUE) and discovered that L. racemosa was not able to increase foliar levels of N even after sites were fertilized with N and P because of low PWUE.Whereas our PNUE data indicate codominance of the L. racemosa and R. mangle regeneration pool at high salinity and higher N exchange and dominance of L. racemosa over R. mangle at high salinity and low N exchange, coexistence of L. racemosa with Avicennia germinans was predicted to occur with low N fertility at moderate salinities [18].

Leaf Gas Exchange and Tidal
Flooding.For most plants, short-term flooding causes acute reductions in leaf gas exchange [54].Prolonged flooding can induce persistently depressed overall physiological activity and growth in flood sensitive species [55,56].Anaerobic soils, as a consequence of flooding, have been associated with decreased   and depressed water potential in flood tolerant mangrove leaves as well [57][58][59].Except under extremely reduced soils (e.g., [59]),   and  appear unaffected by permanent flooding of R. mangle and L. racemosa seedlings or saplings [31,59,60], even though prolonged flooding and/or increases in flood depth can prompt changes in other types of physiological parameters (e.g., root oxidase activity, expression of chlorophyll pigments [61,62]).Mangrove leaf gas exchange, thus, has a capacity to function under a range of hydroperiods, as long as flooding does not significantly lower soil redox conditions or, as we show here, create different salinity conditions.
Redox state ranged from −112 to 149 mv in this study, commensurate with a range of −212 to 292 mv for field measurements from the literature [21,[63][64][65][66].While we were surprised that there were no consistent differences in leaf gas exchange for tidal versus flooded hydroperiods in our study, similar redox conditions (sensu [32]) suggest that perhaps our 12 h day −1 simulated tidal hydroperiod did not differ enough from permanent flooding to stimulate a critical disparity in soil condition.Ultimately, interactions between salinity and hydroperiod will need to be directed at a wider range of hydroperiods in order to simulate early growing conditions of these Neotropical mangroves.

Conclusions
From the perspective of leaf-level gas exchange, it is difficult to understand why R. mangle and L. racemosa have different field preferences for salinity and hydroperiod unless we expand on our understanding of specific nutrient and water use relationships.It is especially important to consider whether PNUE also decreases with increases in salinity without regard to hydroperiod in the field as in the greenhouse (Figure 4), as evidence does suggest that interactions between soil condition and hydroperiod affect relationships among supply, plant demand, and root acquisition of nutrients [67].The coexistence of L. racemosa and A. germinans is facilitated by shifts in PNUE and PWUE with edaphic fertility [18].We suggest here that similar among-species partitioning at the leaf level with shifts in salinity can explain the success of L. racemosa versus R. mangle seedlings in some Neotropical floodplain locations through the differential ability to use nitrogen and water efficiently in photosynthesis on sites with contrasting N fertility.In contrast, differences in PNUE and PWUE imposed by hydroperiod alone explained very little.Including tidal flood frequencies and durations more indicative of upper intertidal locations (lower flood durations) will improve our overall understanding of flooding effects across the floodplain.
International Journal of Forestry Research Mancera, and Edward Castañeda-Moya for assistance with field logistics, research ideas, and maintenance of greenhouse experiments.Luzhen Chen, Rebecca J. Howard, and Darren J. Johnson provided helpful peer, statistical, and editorial reviews of past paper drafts.Experiments were supported by the National Science Foundation under Cooperative Agreement no.DEB-9910514 (Florida Coastal Everglades, Long-Term Ecological Research) and by the University of Louisiana at Lafayette Center for Ecology and Environmental Technology.Instituto Colombiano para el Desarrollo de la Ciencia y Tecnología (Colciencias) provided a scholarship to the first author.Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the US Government.

FFigure 5 :
Figure 5: Stomatal conductance (  ) as a function of the relative humidity within the sample cuvette of the Li-6400 for Laguncularia racemosa and Rhizophora mangle seedlings during two time periods.Data were pooled across species.

Table 1 :
Mixed model ANOVA results describing leaf gas exchange variables for seedlings of Rhizophora mangle and Laguncularia racemosa grown in mixed culture under different salinity and hydroperiod treatments.