The Nitrogen Paradox in Tropical Forest Ecosystems

Observations of the tropical nitrogen (N) cycle over the past half century indicate that intact tropical forests tend to accumulate and recycle large quantities of N relative to temperate forests, as evidenced by plant and soil N to phosphorus (P) ratios, by P limitation of plant growth in some tropical forests, by an abundance of N-fixing plants, and by sustained export of bioavailable N at the ecosystem scale. However, this apparent up-regulation of the ecosystem N cycle introduces a biogeochemical paradox when considered from the perspective of physiology and evolution of individual plants: The putative source for tropical N richness—symbiotic N fixation—should, in theory, be physiologically down-regulated as internal pools of bioavailable N build. We review the evidence for tropical N richness and evaluate several hypotheses that may explain its emergence and maintenance. We propose a leaky nitrostat model that is capable of resolving the paradox at scales of both ecosystems and individual N-fixing organisms. 613 A nn u. R ev . E co l. E vo l. Sy st . 2 00 9. 40 . D ow nl oa de d fr om a rj ou rn al s. an nu al re vi ew s. or g by U ni ve rs ity o f C al if or ni a Sa nt a B ar ba ra o n 11 /1 0/ 09 . F or p er so na l u se o nl y. ANRV393-ES40-29 ARI 9 October 2009 12:38 WHAT SUSTAINS NITROGEN RICHNESS IN TROPICAL FORESTS? Tropical forests play a pivotal role in regulating Earth’s climate and biogeochemical cycles through their vast exchanges of energy, water, carbon, and nutrients with the global atmosphere (Bonan 2008, Brown & Lugo 1982, Cleveland et al. 1999, Field et al. 1998, Melillo et al. 1993). Because of their high productivity, tropical forests account for a large proportion of global terrestrial carbon storage and cycling and are thought to play a critical role in buffering the atmosphere against rising CO2 (Chave et al. 2008, Gerber et al. 2009, Lewis et al. 2009, Malhi & Phillips 2004, Phillips et al. 1998). Yet, despite the well-recognized importance of this biome to biogeochemistry as well as biological diversity, fundamental uncertainties remain in our understanding of the nutrient cycles that underlie the productivity and dynamics of these forests worldwide. One of the most vexing problems lies in the resolution of the nitrogen (N) cycle across this vast biome. Although tropical forests are quite variable in biotic composition and functional properties, it is often assumed that humid lowland tropical forests generally are rich in N relative to other nutrients such as phosphorus (P) or calcium (Ca). This assumption is supported by evidence indicating that at least some tropical forests possess the capacity to build up, recycle, and export (via leaching and denitrification) very large quantities of N (e.g., Davidson et al. 2007, Hall & Matson 1999, Hedin et al. 2003, Houlton et al. 2006, Jenny 1950, Martinelli et al. 1999, Vitousek 1984, Vitousek & Sanford 1986). This capacity introduces a stark and fundamental contrast in our conception of forests as biogeochemical systems: Temperate forests are seen as subject to strong and persistent N limitation, whereas lowland tropical forests are viewed as possessing the exceptional capacity to develop abundant supplies of N. Is such a biome-scale dichotomy in N cycling real and, if so, what maintains it? The view of lowland tropical forests as N rich is subject to two fundamental problems. First, the generalization is based on limited empirical information that often is indirect and nonexperimental. Second, and more fundamentally, it is theoretically difficult to resolve the emergence of N richness at the ecosystem level, based on mechanisms that operate at the level of physiology and ecology of individual N-fixing organisms. The most commonly offered explanation—that biological N fixation (BNF) brings in large amounts of new N from the atmosphere (e.g., Cleveland et al. 1999, Crews 1999, Galloway et al. 2004, Jenny 1950, Robertson & Rosswall 1986, Vitousek & Howarth 1991)—suffers from the problem that, at the organismal level, BNF ought to be down-regulated in N-rich environments (e.g., Barron 2007, Hartwig 1998, Menge et al. 2009a, Pearson & Vitousek 2001) and therefore cannot be invoked as a mechanism that can sustain richness in N over other resources (Hedin et al. 2003). Hedin et al. (2003) proposed that buildup of bioavailable N in humid tropical forests presents a major unresolved paradox in the terrestrial N cycle with implications for understanding how organisms function within ecosystems and how forests function within the global Earth system. We here review the evidence for the idea that humid tropical forests can naturally develop high bioavailability of N relative to other resources. We also examine three competing mechanisms that individually, or in combination, can resolve this paradox of N richness in tropical forest ecosystems. Finally, we discuss some key challenges for the path toward better understanding of the N cycle in this important biome. THE TROPICAL NITROGEN PARADOX In 1950, Hans Jenny first raised the question of why some tropical forests appear to build up and recycle very large amounts of N ( Jenny 1950). Based on the limited data available at the time, and using back-of-the-envelope calculations of soil nutrient turnover ( Jenny et al. 1948, 1949), 614 Hedin et al. A nn u. R ev . E co l. E vo l. Sy st . 2 00 9. 40 . D ow nl oa de d fr om a rj ou rn al s. an nu al re vi ew s. or g by U ni ve rs ity o f C al if or ni a Sa nt a B ar ba ra o n 11 /1 0/ 09 . F or p er so na l u se o nl y. ANRV393-ES40-29 ARI 9 October 2009 12:38 Jenny inferred that humid tropical forests must depend on external N inputs that are substantially greater than those observed in temperate regions. What factor, asked Jenny, can sustain such large N inputs to tropical ecosystems, when conventional wisdom holds that N is an exceedingly rare resource that limits plant growth in many terrestrial ecosystems? Jenny invoked a biological mechanism to explain this pattern: Symbiotic N-fixing plants— primarily in the family Fabaceae (Leguminosae)—act to bring in large amounts of new N from the atmosphere. Specifically, he argued that because symbiotic N fixers are both abundant and diverse in the tropics (but not in temperate regions), this unique plant group could supply the external N needed to sustain rich N availability at the ecosystem level. We have learned a great deal about tropical nutrient cycles in the six decades following Jenny’s seminal observations. We now know that the N content of soils varies greatly across terrestrial ecosystems worldwide, with relatively high stores occurring in moist and productive environments such as tropical and subtropical forests (Post & Pastor 1985). We also know that tropical forests can, over centuries to millennia of ecosystem development, become exceedingly rich in N relative to P and/or other plant resources (Harrington et al. 2001, Hedin et al. 2003, Herbert & Fownes 1995, Vitousek & Farrington 1997). Studies of tropical soils have similarly shown that N can build up to levels presumed to be in excess of other mineral resources, most notably P and base cations such as Ca (Cleveland et al. 2002, Crews et al. 1995, Fox et al. 1991, Hall & Matson 1999, Van Wambeke 1992, Wardle et al. 2004). In addition, recent observations show that some tropical forests appear to sustain substantial losses of bioavailable forms of N (Hedin et al. 2003, Houlton et al. 2006, Lewis et al. 1999) at rates that exceed those thought to be diagnostic only of highly polluted and N-saturated temperate forests. Finally, efforts to construct regionalor global-scale N budgets have generally accepted the presumption that BNF rates are high in tropical forests compared to their temperate, boreal, or arctic counterparts (Cleveland et al. 1999, Galloway et al. 2004, Robertson & Rosswall 1986). It should be noted, however, that these budget calculations derive from only a few field studies of BNF and other N fluxes, extrapolated to vast regions of the forested tropics. Although these observations generally support the notion that N richness can develop in tropical forests, they raise two serious problems relative to Jenny’s interpretation that this phenomenon is caused by high rates of symbiotic BNF. First, the field observations come from few studies in few locations, such that our knowledge of tropical nutrient cycles is geographically selective. This is particularly germane as tropical forests occupy an extraordinarily broad range of variation in biotic and abiotic factors including climate, biota, geological parent material, relief, soil conditions, and age and frequency of disturbances. Rather than being monolithic and static, these forests likely display considerable variation in plant-nutrient-soil dynamics, calling for careful use of abstraction and generalization (e.g., Grubb 1977; Herrera et al. 1978; LeBauer & Treseder 2008; Townsend et al. 2007, 2008; Vitousek 1984). We here review the strengths, limitations, and implications of the present evidence for N richness in tropical forests. A second and conceptually deeper issue concerns the apparent paradox that is the central theme of this review: N fixation in the face of N richness (or, vice versa: N richness in the face of costly N fixation). While Jenny’s idea of symbiotic BNF can help explain the emergence of N richness at the level of ecosystem dynamics, it fails to explain why fixation should be maintained physiologically, ecologically, and evolutionarily at the level of individual plants. Given the substantial energetic cost of BNF (Gutschick 1981), and given that N fixers must remain competitive against nonfixing plants, fixers ought to down-regulate fixation once N availability builds up in excess of demand in the local environment. Such facultative fixation is well recognized experimentally, in symbiotic fixers (Ingestad 1980, Olsen et al. 1975, Pate & Dart 1961) as well as in nonsymbiotic (heterotrophic) Nfixing bacteria in soils (Barron et al. 2009, Crews et al. 2000). The paradox of N richness therefore www.annualreviews.org • Nitrogen in Tropical Forests 615 A nn u. R ev . E co l. E vo l. Sy st . 2 00 9. 40 . D ow nl oa de d fr om a rj ou rn al s. an nu al re vi ew s. or g by U ni ve rs ity o f C al if or ni a Sa nt a B ar ba ra o n 11 /1 0/ 09 . F or p er so na l u se o nl y. ANRV393-ES40-29 ARI 9 October 2009 12:38 identifies a fundamental contradiction across scales of biological organization: Though BNF can sufficiently resolve the emergence of N richness at the level of ecosystems, the explanation is inconsistent with theoretical expectations of how BNF should function as an adaptive strategy at the level of individual organisms. McKey (1994) advocated a slightly different perspective by suggesting that the ecological benefits associated with the N-rich lifestyle of legumes (many of which are N fixers)—such as elevated photosynthetic capacity, N-rich seeds, or increased herbivore defense by N-rich compounds— may outweigh the costs of BNF even in habitats where soil N is sufficient for nonfixing plants. This perspective offers support for the idea that N fixers fix even when N availability is high and, as a result, fixation can be the major determinant of N-rich conditions in tropical lowland forests (Cleveland et al. 1999, Houlton et al. 2008, Jenny 1950). The perspective introduces the idea of overfixation, which we here define as fixation that is sustained even in the face of high N availability. Such overfixation can, in theory, maintain N inputs at the ecosystem scale, even beyond what is needed to alleviate N limitation. Overfixation can result if fixers are physiologically unable to reduce BNF in environments with high bioavailable N. Obligate fixation is the, perhaps, most pure form of overfixation, in which fixers are incapable of adjusting BNF in response to any change in local levels of bioavailable N. Overfixation can also result from an imperfect feedback between BNF and N availability, such that down-regulation of BNF occurs only once N has accumulated well beyond N limitation. Although differing in magnitude, imperfect regulation and obligate fixation yield the same effect: maintenance of N fixation beyond what is needed to alleviate N limitation. In a competitive environment, such overfixation is likely energetically costly compared to the strategy of instantaneous, facultative down-regulation. It is therefore reasonable to expect strong selective pressure on the ability of individual fixers to fine-tune BNF to match local N availability, to the extent such fine-tuning is possible (Menge et al. 2008). Using a simple ecosystem model, we next explore the influence of three alternative fixation strategies on the buildup and loss of bioavailable N at the ecosystem level: absence of N fixation, obligate overfixation (constant N fixation per unit plant biomass), or perfect facultative fixation (down-regulation of BNF once N limitation is alleviated). We then compare this analysis against empirically observed patterns of N cycling. INFLUENCE OF FIXATION DYNAMICS ON ECOSYSTEM NITROGEN RICHNESS Inputs of external N by fixation can have complex and even counterintuitive effects on the availability, cycling, and loss of N at the ecosystem level. Of particular concern is whether a dynamic feedback is established between fixation and local N availability. We here summarize the consequences of different fixation dynamics by showing simulation results from a plant-soil ecosystem model. Our analysis builds on earlier models (Menge et al. 2008, 2009a,b) by incorporating a second mineral nutrient resource in addition to N and allowing different dynamics of BNF. Although the second resource could be any growth-limiting nutrient, we here consider P because it is often assumed to limit plant growth in many tropical soils (Porder et al. 2007, Vitousek 1984). As illustrated schematically in Figure 1, we let plant biomass growth (dB/dt) depend on uptake of bioavailable N (AN ) and P (AP) from the soil and from N fixation (F), and we express the cost of N fixation as a reduction in net plant growth: