ISOTOPIC VIEW OF VEGETATION AND CARBON AND NITROGEN CYCLES IN A CERRADO ECOSYSTEM, SOUTHEASTERN BRAZIL

Carbon and nitrogen biogeochemical cycles in savannas are strongly regulated by the seasonal distribution of precipitation and pulses of nutrients released during the wetting of the dry soil and are critical to the dynamics of microorganisms and vegetation. The objective of this study was to investigate the spatial and temporal variability of C and N isotope ratios as indicators of the cycling of these elements in a cerrado sensu stricto area, within a protected area in a State Park in the state of São Paulo, Brazil. The foliar δC and δN values varied from –33.6 to –24.4 ‰ and –2.5 to 4.5 ‰, respectively. The δC values showed a consistent relationship with canopy height, revealing the importance of structure of the canopy over the C isotopic signature of the vegetation. Carbon isotopic variations associated with the length of the dry season indicated the importance of recent fixed C to the integrated isotopic signature of the leaf organic C. The studied Cerrado species showed a depleted foliar δN, but a wide range of foliar Nitrogen with no difference among canopy heights. However, seasonal variability was observed, with foliar δN values being higher in the transition period between dry and rainy seasons. The variation of the foliar C and N isotope ratios presented here was consistent with highly diverse vegetation with high energy available but low availability of water and N.


INTRODUCTION
Savanna ecosystems are primarily controlled by the interactions between water and nutrient availability, as savanna vegetation often occurs either on highly weathered soils with low nutrient availability, in regions of limited rainfall, or both (Medina, 1987).This basic environmental structure can be modified by changes in fire frequency and land-use (Bustamante et al., 2006;Miranda et al., 2002;Ratter et al., 1997).
The Brazilian Cerrado (savanna) biome covers 2 million km 2 representing 23% of the area of the country.It is an ancient biome with rich biodiversity, with approximately 800 species of trees and large shrubs (Ratter et al., 1997).Despite the intense land-use conversion of the native Cerrado in the recent decades, few studies aim to understand functioning of the Cerrado ecosystem.Furthermore, studies on the influence of vegetation on carbon and nitrogen cycles in this savanna ecosystem remain scarce.
The C and N dynamics are primarily modulated by the biota, especially through photosynthesis as one of the major links between atmosphere and terrestrial ecosystems.To better understand the C and N cycles and how they affect the ecosystem functioning, it is essential to understand how plants interfere in the C and N cycle in the ecosystem.One way to address this issue is to use integrating tools such as the natural abundance of C and N isotopes.While several previous studies have focused on qualitative patterns of nitrogen stable isotope ratios (δ 15 N) in ecosystems to derive information about the N cycle (Amundson et al., 2003;Bustamante et al., 2004;Martinelli et al., 1999;Ometto et al., 2006), variation in the carbon stable isotope ratios (δ 13 C) allows for making inferences about physiological plant dynamics, but also the ecophysiological aspects of the ecosystem (Ometto et al., 2006;Pataki et al., 2003).
The objective of this study was to investigate the spatial and temporal variability of C and N isotope ratios as indicators of the cycling of these elements in a cerrado sensu stricto area, in São Paulo State, southeastern Brazil.

Study site
The studied site is located in a state park in the municipality of Santa Rita do Passa Quatro (21°36-38'S and 47°36-39'W), in the State of São Paulo, Brazil.The park has 2,070 ha, with altitude varying from 590 to 740 m.The cerrado vegetation covers 59% of the park, with the cerrado sensu stricto being the predominant phytophysiognomy (79%) (Batalha & Mantovani, 2001), composed of a continuous ground layer dominated by grasses and a woody layer of trees and shrubs covering 50-60% of the area, with maximum canopy height of 10m (Pivello et al., 1998).This area has been protected against fire since it was converted to a state park in 1970 (H.J. Ribeiro, personal communication).From the total Cerrado area in Brazil, about 6% occur in the state of São Paulo, representing the southeastern boundary of the Cerrado biome; however, this remaining natural vegetation appears nowadays as small fragments of the landscape, mostly surrounded by eucalyptus, citrus and sugar cane plantations (Vieira et al., 1989).

Plant and Soil Sampling
The foliar material was collected during a two-year period (from May 2005 to April 2007).A total of 180 leaf samples were collected throughout the canopy height profile, using an 18 m scaffold tower.For each individual plant, five to eight mature and healthy leaves were sampled from the same height.A composite sample was made with leaves sampled at the same height from the same individual plant.All sampled trees were positively identified, and the plants belonged to 22 Cerrado wood species.The plants were grouped according to the corresponding canopy height and classified as: understory (≤ 2 m); mid canopy (> 2 m and < 7 m) and upper canopy (≥ 7 m).Four soil samples were taken at the soil surface (0-10 cm depth) near the scaffold tower mentioned above.

Isotopic analyses
All leaf samples were oven-dried at 65°C until constant weight and then ground to a fine powder.Soil samples were air-dried, sieved using a 2-mm mesh and homogenized.A smaller sub-sample was taken, handpicked to remove fine roots and other debris and then ground in a mortar and pestle.
A 1.5-2 mg sub-sample of ground leaf material or 15-20 mg sub-sample of ground soil were placed and sealed in a tin capsule and loaded into a ThermoQuest-Finnigan Delta Plus isotope ratio mass spectrometer (Finnigan-MAT; CA, USA) in line with an Elemental Analyzer (Carlo Erba model 1110; Milan, Italy).Stable isotope ratios of C and N were measured relative to recognized international standards.Internal working standards (Atropine and soil standard no.502-308 from LECO Corporation) were included in every run, as a standard laboratory procedure.Stable isotope values are reported in "delta" notation, as δ values in parts per thousand (‰), so that δ‰ = (R sample / R standard -1) × 1000, where R is the molar ratio of the rare to abundant isotope ( 15 N/ 14 N; 13 C/ 12 C) in the sample and the standard.The precision of the measurements were ± 0.3%, 0.1%, 0.3‰ and 0.5‰ for C, N, δ 13 C and δ 15 N, respectively.

Statistical Analyses
The Kolmogorov-Smirnov test was used to determine the normality of the data.As the data followed normal distribution, parametric tests and Pearson correlations were used.Student's t-test was used to assess differences between legumes and non-legumes within a site, or for seasonal comparisons both for C and N data.ANOVA followed by a post hoc Tukey HSD test was used to determine differences among canopy heights.All statistical analyses were performed using the software STATISTICA, version 6.1 for Windows (StatSoft, 2004).Differences at the 0.05 level were reported as significant.

Soil characterization
The C and N isotope ratio of the studied soil (0-10 cm depth) was -25.7 ± 0.9 ‰ and 4.2 ± 0.9 ‰ (mean ± standard deviation), respectively.The total soil N concentration was 0.6 ± 0.2 g kg -1 and the C:N ratio was 16 ± 2.0.

Overall tendencies
The foliar δ 13 C and δ 15 N values varied from -33.6 to -24.4 ‰ and -2.5 to 4.5 ‰, respectively.There were no relationships between foliar δ 13 C and δ 15 N values (r = -0.06).The species average for foliar δ 13 C, δ 15 N, N concentration and C:N ratio are presented in Table 1.

Height variability of the foliar isotope values
Regardless of the period sampled, foliar δ 13 C values were positively correlated with canopy height (r = 0.66; p < 0.05) (Figure 1) while foliar δ 15 N values did not correlate with canopy height (r = -0.08).In general, individuals belonging to the upper and middle canopy showed no differences in foliar δ 13 C averages, but a pattern of higher foliar δ 13 C could be observed for those individuals in the upper canopy (Figure 2).On the other hand, individuals in the understory showed consistently lower values of foliar δ 13 C in comparison to those of the other canopy heights (p < 0.05) (Figure 2).In terms of foliar δ 15 N, no differences were found among individuals from all canopy heights.Therefore, in the following sections, all the foliar δ 15 N data include individuals from all canopy heights, unless specified.

Stable isotope ratios and nitrogen content of the woody plants
The studied species were separated according to their putative N 2 -fixing ability (individuals belonging to the Leguminosae family and hereafter called legumes) or inability (individuals belonging to families other than legumes and hereafter called non-legumes) (Table 1).Considering all samples, the foliar δ 13 C average values of legume leaves were higher than the non-legume leaves while mean foliar δ 15 N values did not differ between legumes and non-legumes.However, the observed range of foliar δ 15 N values for the legumes was notably smaller (-1.2 to 1.4 ‰) compared to the variation found for the non-legume species (-2.5 to 4.5‰).
The total average foliar N concentration was 2.7 ± 0.5% for legumes, which was higher than the 1.8 ± 0.4% for the non-legume trees (p < 0.05).Foliar C:N ratio was, on the other hand, higher for the non-legume (27.8 ± 7.5) than for the legume trees (19.8 ± 3.5) (p < 0.05).
Despite an absence of difference in foliar δ 15 N between legumes and non-legumes, the legume species that had at least three individuals samples in this study were classified into two groups: potentially N 2 -fixing legumes and non N 2 -fixing legumes, according to the literature (Table 1) and were compared to two nonlegumes (Figure 3) to determine whether they were fixing N 2 from the atmosphere.There was an overlap among the foliar δ 15 N values of the potentially N 2 -fixing legumes, the non N 2 -fixing legumes and the nonlegumes, and it was not possible to identify their ability to fix atmospheric N using natural abundance of 15 N.

Temporal variability
The sampling strategy was devised to investigate the temporal variability of the δ 13 C and δ 15 N values of tree leaves over a two-year period (Figure 2 and 4).When the data were classified as dry season (May to September) and rainy season (October to April), no differences were found for δ 13 C in tree leaves from all canopy heights.However, there was a tendency for the foliar δ 13 C signals of the upper and middle canopy trees to increase during the transition period between the dry and rainy seasons, which seemed to follow a one-year cycle.The lowest values occurred in the transition from the rainy to dry seasons (Figure 2).The increase of water availability in the system can poten- tially change the δ 13 C of the leaves changing the ci/ca ratio (see Farhquar et al., 1989), although the patterns observed for upper and middle canopy trees indicated a difference between the expected isotopic signature of recent assimilated C and the total leaf C which reflects the integrated C isotopic signature.For the understory species this tendency was not notable during the sampled years.
The same seasonal grouping of the data showed no differences for δ 15 N in tree leaves, which can be related to the high variability in the foliar δ 15 N among the studied species (Table 1).However, there was a notable response in δ 15 N values to variations in precipitation, with higher δ 15 N values occurring in the transition period between dry and rainy seasons and lower values occurring in the transition between rainy and dry seasons (Figure 4).This seasonal tendency was observed in both sampled years, but a variation in the magnitude on δ 15 N values could be noted, both for legume and non-legume species.
Table 1 -Mean and standard deviation of δ 13 C and δ 15 N, foliar N concentration, and C:N ratios in leaves of Cerrado tree species classified in three canopy heights: understory, middle canopy and upper canopy.

DISCUSSION
The data presented here indicate a high variability of the foliar δ 13 C among the Cerrado species but strongly associated with canopy height.The overall average δ 13 C value for all leaf samples (-30.2 ± 1.8 ‰; n=180) (mean ± 1 SD) was higher than the average values reported for tropical rainforests (Bonal et al., 2000;Buchmann et al., 1997;Ometto et al., 2006).Medina & Minchin (1980) and Ometto et al. (2006) showed a significant C isotopic variation in relation to the pho-tosynthetically active radiation (PAR) between the upper and lower part of an Amazonian terra-firme forests, named the "canopy effect".For the dense canopy of the Amazonian terra-firme forests, for example, plants located in lower canopy strata had a different distribution and quantity of stomata, compared to the upper canopy leaves.
The lower incidence of direct radiation and possible low atmospheric water vapor deficit (VPD), associated with greater stability of the microclimate in this section of the canopy, allows the understory plants to maintain their stomata open for longer periods (Ometto et al., 2002;Ometto et al., 2006).Hence, leaves in the upper part of the canopy strata discriminate less in regards to carbon 13, reflecting in heavier δ 13 C values.In this study, the differences in the δ 13 C values found for the canopy strata suggest different strategies of plants for assimilating CO 2 , as also observed by Ehleringer et al. (1993).As briefly mentioned above, the response of photosynthesis to precipitation or water availability will be reflected in the carbon isotope ratios of the vegetation (Farquhar et al., 1989).The C isotope ratios of foliar tissue represent a balance between stomatal conductance to supply CO 2 and photosynthetic demand consuming CO 2 , integrated in the ci/ca ratio.In the Brazilian Cerrado, little data on the C isotopic variation has been published.Mattos et al. (1997), looking at the δ 13 C of 13 species from 11 families, in a Cerrado area near this study area (São Carlos, SP) calculated the intrinsic water use efficiency and identified limitations to carbon fixation caused by water stress.
The strong seasonality observed in precipitation and water availability was not clearly shown in our δ 13 C values, although, in the upper canopy strata, plants seem to be more sensitive to variation in water availability.In the Amazonian terra-firme forests the seasonal distribution of precipitation is correlated with the δ 13 C variability of tree leaves (Ometto et al., 2006).
Their observations suggested a time lag between the C structurally allocated on the leaves' tissues and the C fixed by photosynthesis, being correlated, therefore, with previous rainfall.In general, a trend indicating a lighter δ 13 C signal with an increase of water availability and ci/ca ratio is typically observed (Walcroft et al., 1997).Some authors working in the Amazonian terra-firme forests with gas flux studies (eddy-covariance) and biomass measurements observed small wood increments during the dry season, whereas a significant increase was observed just before the start of the rainy season (Goulden et al., 2004;Saleska et al., 2003), which could indicate a lighter isotope signal of C fixed at the beginning of the rainy season.The mechanism suggested by Saleska et al. (2003) for the increase of diameter wood increment immediately preceding the return of the rain could be related to the specific C re-allocation by the plant.
The higher foliar δ 13 C of middle and upper canopy during the dry season might reflect the cerrado sensu stricto canopy structure, straining this vegetation in a higher VDP atmosphere and the lower water availability, typical of the season.The range of the foliar δ 13 C data of the lower canopy individuals is similar to the lower canopy δ 13 C data of the Amazonian terra-firme forests throughout the year.This seems to reflect an adaptation of the species located in this stratum to a more stable micro-climate, discriminating against the heavier carbon ( 13 C), therefore reflecting a higher relationship between stomata conductance and photosynthetic capacity.
Variations in δ 13 C values have been correlated with other parameters not directly related to canopy height, including water-use efficiency (Bonal et al., 2000;Guehl et al., 1998), branch morphology and hydraulic conductivity (Panek, 1996), structural and compositional differences between leaves (Bonal et al., 2000;Broadmeadow & Griffiths, 1993), and leaf age and position in the crown (Gebauer & Schulze, 1991).However, the relative contribution of these factors to variations in δ 13 C of plant tissue remains unknown.
Plant species in the studied cerrado sensu stricto area showed depleted foliar δ 15 N signatures, and there was no difference between canopy heights.However, the range of foliar δ 15 N was similar to the wide variation in foliar δ 15 N reported by Bustamante et al. (2004) in different areas of cerrado sensu stricto in Central Brazil.On the other hand, data from Brazilian forest ecosystems have consistently higher foliar δ 15 N with a narrow range of values compared to the Cerrado data (Aidar et al., 2003;Ometto et al., 2006).Several factors contribute to the large variability in foliar δ 15 N in the Cerrado, including associations with mycorrhizal fungi, the variability in the levels of δ 15 N in the soil organic matter according to depth, the seasonality of N-immobilization and N mineralization (Bustamante et al., 2004).Despite the large foliar δ 15 N variability among the Cerrado trees, seasonal variability was observed in the N isotopic composition of the Cerrado plant leaves, with foliar δ 15 N values increasing during the transition period from the dry to the rainy seasons.Some factors may contribute to the seasonal pattern found, as discussed in detail by Bustamante et al. (2004), including the fact that older tissues tend to have been higher δ 15 N than younger ones due to the translocation of N, suggesting that the net isotopic effect of the phenological cycle in the Cerrado would lead to be higher δ 15 N values during the dry season.Whole-plant and leaf N isotope composition are determined both by the isotope ratio of the external N source and physiological mechanisms within the plant; however, whole-plant N isotope composition can reflect that of the N source when plant demand exceeds N supply (Evans, 2001).
In the cerrado sensu stricto, although net N mineralization and net nitrification rates are higher during the rainy season, inorganic-N concentrations decrease during the rainy season, indicating that plant uptake decreases inorganic-N stocks and consumes the N released from net mineralization (Nardoto & Bustamante, 2003).Despite the transient nature of nitrate in soils, the low content of NO 3 -N in soils of the Cerrado sites could be related to low nitrification rates, even though significant production of NH 4 -N occurs.
The rate of net N mineralization in a cerrado sensu stricto area in Central Brazil protected from fire for 28 years was only 14.7 kg N ha -1 yr -1 and NO 3 -N accumulation occurred only in short periods during the rainy season (Nardoto & Bustamante, 2003).In addition, the high C:N of litter (>50:1) may contribute to the low rates of decomposition and mineralization in Cerrado areas, thus maintaining low availability of N in this ecosystem (Bustamante et al., 2006).
Mechanistic explanations for local or regional patterns of δ 15 N values would require the use of Rayleigh equations to determine variations in the product (e.g.plants) and the source (e.g.nitrogen) (Robinson, 2001).This equation establishes the relationship between the isotopic composition of precursor and product based on the fractionation factor and change in concentration of the substrate.Although, according to Högberg (1997), the complexity of N dynamics in the plant-soil system renders it quite difficult to apply a source-product isotopic model.Nonetheless, analysis of the 15 N natural abundance can provide information on sources and transformations of N in ecosystems (Amundson et al., 2003;Högberg, 1997;Martinelli et al., 1999;Roggy et al., 1999), since variations in δ 15 N reflect N isotope fractionations where, for a given reaction, the product typically becomes 15 N depleted relative to the substrate (Högberg, 1997).
We did not measure the isotopic composition of soil nitrate; however, there is significant fractionation during the nitrification process, depositing in the soil, isotopically depleted nitrate as compared to the ammonium (Högberg, 1997).Consequently, nitrate uptake during the rainy season would result in plants with lower δ 15 N values in comparison to the dry season when ammonium predominates in the cerrado sensu stricto soil (Nardoto & Bustamante, 2003).Therefore, isotopically identifying the N source of Cerrado woody plants remain an open prospect.Högberg (1997) predicted that if plants derived a substantial portion of N inputs from atmospheric N 2 , then leaf δ 15 N values would be approximately 0-2 ‰.The δ 15 N method might be inadequate for estimating N 2 fixation in such N-limited conditions where plants tend to be depleted in 15 N as observed in the Cerrado plants because the foliar δ 15 N values of the legumes overlapped with those of the non-legumes.Nonetheless, nitrogen biological fixation is considered an important process for delivering N to the system, especially since this element is limited and there is enough light to induce biological fixation.Gehring et al. (2005) considered significant symbiotic biological nitrogen fixation (symbiotic BNF) in secondary regrowth (where N-limiting conditions tend to be higher) contrary to low symbiotic BNF in mature rainforest in Central Amazon.Given that areas with N-limited conditions such as the cerrado sensu stricto studied herein, the role of symbiotic BNF is supposedly an important N input to the system.Bustamante et al. (2006) suggest that most Cerrado legumes, particularly those of the subfamily Faboideae, are active N 2 -fixers under natural conditions, but for a reliable assessment of the contribution of legumes to the N budget of savannas, it is necessary to understand controls over legume density and seasonal nitrogen fixation activity.However, regardless of whether they are active N 2 -fixers or not, the studied legume species showed consistently higher foliar N concentration than the non-legume.This same overall trend has been observed for other Cerrado areas (Bustamante et al., 2004;Sprent et al., 1996), and tropical rainforests Sci.Agric.(Piracicaba, Braz.), v.66, n.4, p.467-475, July/August 2009 (Ometto et al., 2006;Roggy et al., 1999;Vitousek et al., 2002), which may be related to a more efficient N assimilation process in legumes with a high-N-demanding character (McKey, 1994).
The difference in canopy structures is an important factor in the ecological understanding of the system, and the δ 13 C values obtained in this study showed a consistent response, changing its value according to canopy height, reflecting the non-homogeneous cerrado sensu stricto structure.The seasonal variability of rainfall indicates the importance of recent fixed C to the integrated isotopic signature of the leaf organic C and, concerning the N dynamics in the ecosystem, the seasonality of water availability seems to play a major role in the plant uptake pattern of N. Therefore, the variation of the foliar C and N isotope ratios presented here was reliable for highly diverse vegetation having high energy but low availability of water and nitrogen.

Figure 1 -
Figure 1 -Variation in the carbon isotope ratios (δ 13 C) of tree leaves according to canopy height, measured from May 2005 to April 2007.

Figure 2 -
Figure 2 -Temporal variation (mean ± standard error) in the carbon isotope ratios (δ 13 C) of tree leaves at three different heights and average monthly precipitation (gray bars) at a State Park, SP, Brazil: ( -black square) leaves collected from the understory (height ≤ 2 m); ( -open circle) leaves collected from the middle canopy (> 2 m and < 7 m); ( -close triangle) leaves collected from the upper canopy (≥ 7 m).

Figure 3 -
Figure 3 -Leaf nitrogen isotope ratios (δ 15 N) (mean ± standard error) for selected legume and non-legume species which were separated in three groups: (open bars) non-legume species, (black bars) non N 2 -fixing legume species and, (gray bars) potentially N 2fixing legume species.

Figure 4 -
Figure 4 -Temporal variation in the nitrogen isotope ratios (δ 15 N) of tree leaves (mean ± standard deviation), and the data were separated in two groups of plants: (black bars) legume and, (open bars) non-legume individuals.