Leaf and litter nitrogen and phosphorus in three forests with low P supply
Julio Campo·Juan F. Gallardo · Guillermina Hernández
J. Campo (communicating author)
Instituto de Ecología, Universidad Nacional Autónoma de México, A.P. 70-275, Mexico DF 04510, Mexico
Tel.: +52-55-56229027
Fax: +52-55-56161976
e-mail:
J.F. Gallardo
Instituto de Recursos Naturales y Agrobiología, CSIC, AP 257, Salamanca 37071, Spain
G. Hernández
Instituto de Ecología y Sistemática, AP 8029, La Habana 10800, Cuba
AbstractWe compared the N and P contents in the main labile components of the nutrient cycle of three different forest ecosystems (a Mediterranean temperate forest, MTF; a Tropical evergreen forest, TEF; and a Tropical dry forest, TDF) with low P-supply. A mass balance approachwas used to estimate the mean residence times for organic matter, N and P, and examine the flexibility of the N and P intra-system cycling in the three forest ecosystems. We approached this task by combining the published values of the N and P in the foliage, litterfall, forest floor litter, and mineral soil from these three forest ecosystems.Our analysis was consistent with the idea that the leaf (both green and senescent leaves) and litter N increases with increasing temperatures. In contrast, our data did not support the hypothesis that the leaf P decreases with increasing temperatures and precipitation: the leaf P and litterfall P concentrations were higher in the two studied tropical forests than in the temperate forest. The TEF had the highest P concentration of all of the studied forests.The mass balance analysis indicate that, while in TDF P mineralization can run ahead of the stoichiometry of litter decomposition when P is in short supply, that flexibility is much-reduced or absent in TEF and MTF.Although our analysis provide additional evidence suggesting the importance of climatic factors in ecosystem processes in forest ecosystems, it also highlights the role of flexibility in within ecosystem nutrient cycling, specially for P, in ecosystems with low supply of P.
KeywordsMediterranean temperate forest·N:P ratio· Nutrient cycling·Nutrient limitation· Tropical forests
Introduction
Nitrogen (N) and phosphorus (P) are essential for plant metabolism (Lambers et al. 2008), and their restricted availability often limits plant carbon acquisition and growth (Elser et al. 2007; Reich et al. 2010; Fisher et al. 2012). Numerous studies have shown that leaf N and P values reflect the nutrient availability of the soil (Aerts and Chapin 2000). Thus, foliar N and P contents are viewed as indicators of nutrient status that may provide insight into such processes as net primary production. However, the growing focus on ecological stoichiometry (Sterner and Elser 2002) has led to a broader use of the N:P ratios in green leaves to infer any potential nutrient limitations in terrestrial primary productivity (Moraghan et al. 2002; Reich and Oleksyn 2004) and plant growth (Ågren 2008).
The foliage is a metabolically active component of the forest biomass, is turned over rapidly, and, thus, is sensitive to differences in the relative investment of the N and P in the physiological functions of plants within a forest ecosystem (McGroddy et al. 2004). The nutrients that are not reabsorbed by the plants will circulate through the litterfall that constitutes one of the major fluxes of nutrient cycling between plants and soils and, therefore, reflects the constraints on the internal fluxes of N:P on an ecosystem scale. In contrast, the characteristics of standing litter are determined by the interaction of many organisms, suggesting that these biologically controlled N and P pools should vary more between ecosystems than those in the litterfall(Ladanai et al. 2010).
The nutrient contents of senescent-leaf litter, together with climate and soil properties, are strong drivers of decomposition rates and nutrient release (Parton et al. 2007; Manzoni et al. 2008) and, therefore, influence the soil nutrient availability and carbon sequestration. Clear global trends in green-leaf N and P (Reich and Oleksyn 2004; Wright et al. 2004; Niklas et al. 2007; Ordoñez et al. 2009) have been demonstrated. More recently, the global trends in senescent-leaf N and P have been studied as extensively as in their green-leaved counterparts (Yuan and Chen 2009; Hongzhang et al. 2010). However, in contrast to the clear worldwide trends in both the green- and senescent-leaf N and P contents and their ratios, the global patterns of the forest floor litter N and P are not well understood. Given that global climatic changes may influence leaf traits, and ecosystem processes, such as decomposition and nutrient mineralization, it is important to investigate the variation in both green and senescent leaves and the litter N and P in relation to ambient factors to understand whether, and to what extent, the intra-system cycling of N and P may change in the future.To help address this paucity, the present study compared the N and P contents in the main labile components of the nutrient cycle of three different forest ecosystems (a Mediterranean temperate forest, MTF;a Tropical evergreen forest, TEF;and a Tropical dry forest, TDF) with low P-supply. In our study, we also investigated the stoichiometry of the N and P, which is a useful index for understanding the mechanisms of the variation of N and P and for nutrient use strategies (Sterner and Elser 2002; Güsewell 2004; McGroddy et al. 2004; Reich and Oleksyn 2004) and may provide insight into the global carbon cycle.
In general, nutrient limitation in seasonal dry ecosystems (as is the case of the MTF and the TDF) is related to water limitation because dry conditions prevent plant uptake of nutrients from the soil, and reduce the release of nutrients during decomposition. Nutrient limitation also occur in forests on highly weathered, P-poor soils in the humid tropical regions, reflecting that soils can become depleted in P as consequence of P is derived primarily from rock weathering. Evidence from low P supplies for the ecosystem function in the selected TEF and MTF comes from studies of P supply in the soil (Hernández 1999; Turrión et al. 2002). More direct evidence for low supplies of P in TDF was found in studies on seedling and adult tree responses to experimental fertilization (Ceccon et al. 2003; Campo and Vázquez-Yanes 2004).We posed two main questions: (1) do the foliar N and P concentrations reflect broad differences in the nutrient status of the soil? and (2) do the N and P values in the leaves, litterfall, and standing litter vary and reflect differences in the climate (mean annual rainfall, mean annual temperature, and the ratio mean annual rainfall/mean annual temperature) among the three studied forest?
Although climate and nutrient availability regulate net primary productivity and decomposition in all terrestrial ecosystems, the nature and extent of such controls in forest ecosystems remain poorly resolved (see Cleveland et al. 2011). Nitrogen and P cycles differ substantially in their sources and controlling mechanisms (Schlesinger 1997). Vitousek (2004) proposes that some nutrients can be cycled more rapidly within an ecosystem than do other nutrients; and refer to this phenomenon as flexibility. When they are in demand, nutrients with flexible cycles are transferred more rapidly than would be expected based on the underlying stoichiometry of the organisms and/or processes involved. He posits that in an ecosystem the nutrient with the least flexible cycle will ultimately limit biological processes. Here, a mass balance approach was used to examine the flexibility of the N and P intra-system cycling in the three P-poor forest ecosystems. For this, we begin by exploring the mean residence time of organic matter, N, and P in the forest floor, and then discuss their implications to improving our understanding of forest biogeochemistry under conditions of low P-supply.
To address these fundamental questions, we compiled the published values of N, and P in foliage, litterfall, forest floor litter, and mineral soil from three very diverse forest ecosystems. In addition, we also compared the trends and the magnitude of the differences in the N and P contents in the litterfall with those in the standing litter across these three studied forests.
Materials and methods
Selected forest sites
The Tropical dry forest (TDF) site was located just in the DzibichaltúnNational Park, in the southeastern region of the Yucatán Peninsula, Mexico (21º 06’N, 89º 17’ W). The mean annual temperature (MAT) of 25.8 ºC varies by < 6 ºC annually, and the region receives a mean annual precipitation (MAP) close to 760 mm yr-1 (MAP/MAT = 29). Most of the rainfall occurs during the wet season from June to October (summer). The landscape consists of flat areas (less than 10 m above sea level), and the predominant lithology includes late Pliocene material, with numerous areas of exposed limestone. The soils (Lithic rendoll) are mainly shallow (0.05-0.1 m in depth) and rich in organic matter and directly lie over weathered calcium carbonates (Table 1); data for samples collected in the middle of the dry season (March) and rainy season (September). The forest is dominated by deciduous treesand floristicallyLeguminosae is the most important family, with 31% of the species (Ceccon et al. 2002).
The Tropical evergreen forest (TEF) site was located at El Salón (220 49’N, 820 58’W; 400 m above sea level), the nucleus zone of the Sierra del Rosario Biosphere Reserve that belongs to the mountain range of Guaniguanico (the Western part of Cuba), 50 km from Western Havana. The climate in the TEF is wet tropical with an MAP of 2014 mm (most of the rainfall occurs between May and October, with three dry-season months between February and April). The MAT is 24.4 ºC (MAP/MAT = 83). The predominant lithology consists of limestone belonging to the Artemisa Formation from the Higher Jurassic. The soil (Mollic cambisol) is shallow (0.5-1.0 m in depth), with a neutral pH (Table 1); data for samples collected in March (dry season) and in July (rainy season). The forest composition is dominated by Talipariti elatum (Sw.) Fryxell (Malvaceae) and Bahuinia cumanensis HBK(Leguminosae). Other evergreen trees, such as Guarea guidonia (L.) Sleumer (Meliaceae), Zanthoxylum martinicense (Lam.) DC., and Matayba apetala (Macf) Radlk. (Sapindaceae) and climber plants, such as Smilax dominguensis Willd. (Smilaceae), can be also found in the forest community (Herrera et al. 1988).
The Mediterranean temperate forest (MTF) site was located just east of the town of Navafrías, on the southwest end of the Salamanca Province (Western Spain; 40º 2’N, 3º 0’W). The climate in the MTF site is temperate, subhumid Mediterranean, with an MAP of 1580 mm yr-1. The dry season typically occurs between June and September (summer) of each year. The MAT is 11 ºC (MAP/MAT = 144). The landscape consists of a system of hills (Sierra de Gata Mountains, 960 m above sea level). The predominant lithology includes acid schists and slates that are deeply weathered, and the area develops more a weathered (Ortic umbrisol), deeper (-1 m in soil depth) and more acidic soil than its tropical counterparts (Table 1); data for samples collected in August (dry season) and January (rainy season). The forest is dominated by the deciduous Atlantic oak (Quercus pyrenaica Willd., Fagaceae). The MTF vegetation has relatively low diversity but includes a broad representation of different plant life forms, including grasses, shrubs, forbs, and the deciduous oak listed above.
Methods
The methodology was based on the measurements of leaf nutrients (in green- and senescent-leaves), aboveground litterfall nutrients, standing litter nutrients, and nutrient concentrations in the soils. These fundamental data in nutrient cycling (both published and in data bases of contributing authors) referred to these three forest ecosystems are compared among forests.
A detailed description of the nutrient sampling and analysis for the TDF can be found in Campo and Dirzo (2003) and Campo et al. (2007), in Menéndez (1988) and Hernández (1999) for the TEF, and in Gallardo et al. (1998) and Moreno-Marcos and Gallardo-Lancho (2002) for the MTF.
In the three studied forests (i.e. TDF, TEF, and TMF) all the measures were done at mature forests. Nutrient concentrations in green- and senescent-leaves were measured for a one-year period. Mature leavesof the three dominate tree species at TDF and TEF, and ofQ. pyrenaica at TMF were haphazardly retrieved from all parts of the crown to ensure that all zones of the canopy were represented. Senescent leaves of each selected specieswere collected from each of the three sites at weekly intervals over a period of two weeks. These leaves were identified as they had a different color from live leaves (often yellow). Senesced leaves were collected directly from the plants rather than from leaf litter, as we were concerned that decomposition of litter and leaching of leaf nutrients would lead to an underestimate of nutrient concentrations in senescent leaves taken from the soil.
Litterfall data come from monthly collectionsin five litter traps (50 in diameter) per plot, and forest floor litter data from quarterly collections in four circular micro-plots (20 cm in diameter) per plot for three years period. Both, data sets of litterfall and of litter used come from samples collected in the same year. For all sites, litter consists of all dead plant material lying on the forest floor, including the freshly fallen litter and the more finely more decomposed litter fraction. Means of annual litterfall and of annual forest floor litter of four plots were calculated (n = 4 plots) per forest.Annual mineral residence times for organic matter, N, and P were calculated from the ratio of the litter standing crop (or litter nutrient pools) to the annual litterfall production (or litterfall nutrient fluxes).
Data analyses
One way analysis of variance was used to test for differences in characteristics of soil, various leaf, litterfall and litter chemistry, and mean residence time of organic matter and nutrients among forests. Within each forest, differences in mean residence times among organic matter, N, and P were examined using ANOVA. When data were not normally distributed, they were log-transformed before analysis. Multiple comparisons were performed with the Honest Significant Difference (HSD) when statistical differences (P < 0.05) among means were observed.
Results
Table 2 shows the N and P contents in the more labile pools (leaves, litterfall, and standing litter) of the three selected forests. Both the TDF and TEF had higher fresh leaf N and P concentrations than the MTF (by 30 to 38% in the case of N and by 38 to 60% in the case of P), which largely reflected the differences in the N and P requirements between forests with leguminous trees and forests without leguminous trees. Difference among forests in N and P concentrations in senescent leaves largely reflected N and P concentrations in green leaves. Our analysis showed that the studied forests were characterized by a well-constrained N:P ratio in the foliage.
As expected, litterfall production was higher in both of the tropical forests (i.e. the TDF and TEF) than in the MTF (by a factor of 3; Table 2). In addition, the N and P concentrations of the litterfall differed between the tropical and temperate forests (by 21 to 25% in the case of N and by 40% in the case of P). In contrast to the observed trends in the foliage N:P ratios, the N:P ratios in the TDF vs. the TEF were remarkably similar for the litterfall. The Mediterranean temperate forest differed, however, displaying a litterfall N:P ratio that was approximately 30% higher than both of the tropical forests.
The annual N and P returns to the soil differed considerably between the tropical forests and the MTF (by a factor of 4 in the case of N and by a factor of 5 in the case of P). These differences between tropical and temperate forests largely reflected both differences in the litterfall production and in the concentrations of nutrients.
Across all of the studied forests, the TDF had the highest standing litter values (by 3 times), as is show in Table 2. A comparison of the climate data suggested that this broad pattern was primarily caused by variations in the MAP/MAT ratio. There was no obvious difference between forests with the highest MAP/MAT value (i.e. the MTF compared to the TEF). In contrast, both of the tropical forests (i.e. the TDF and TEF) had N concentrations in the standing litter that were four times higher than the MTF. In addition, the TEF had a P concentration in the standing litter that was two times higher than the other two forests (i.e. the TDF and MTF), whereas the latter two forests had litter P concentrations that were similar. It is interesting to note that the N:P ratio for the standing litter increased in a trend with the following order: MTF < TEF < TDF (an opposite trend to the senescent leaves). We also observed a general trend of increased litter N and P pools in the tropical forests. The range of variation across the forests was one fold for the N pool and approximately two fold for the P pool; the lowest minimum litter nutrient pools occurred in the MTF, whereas the highest nutrient pools occurred in the TDF.
We found marked differences in the foliar vs. litterfall N:P ratios across the forests (Table 2). The most conspicuous differences were caused by a divergence in the foliage and litterfall N:P ratios between the TEF (litterfall N:P ratio : leaf N:P ratio = 1.01) and the other two forests (stoichiometric ratio = 1.26 and 1.36, for the TDF and MTF, respectively), indicating a lower tendency for the reabsorption of P relative to N in the TEF, as compared to the other two forests. In addition, the TDF displayed substantially greater differences in the N:P ratios between litter and litterfall (litter N:P ratio : litterfall N:P ratio = 2.21) than the TEF (stoichiometric ratio = 0.95) and MTF (stoichiometric ratio = 0.39), indicating a stronger tendency for the release of P relative to N in the TDF. Interestingly, the N:P ratios in the TEF were remarkably similar for the foliage, litterfall and litter.