Controls Over Leaf Litter and Soil Nitrogen Fixation in Two Lowland Tropical Rain Forests

Global com parisons suggest that rates o f N fixation in tropical rain forests m ay be am ong the highest on earth. However, data supporting this contention are rare, and the factors th a t regulate N hxation w ithin the biom e rem ain largely unknow n. W e conducted a full-factorial (N X P) fertilization experim ent in two lowland tropical rain forests in C osta Rica to explore the effects o f nu trien t availability on rates o f free-living N hxation in leaf litter and soil. P fertilization signihcantiy increased N hxation rates in both leaf litter and soil, and the effect was dependent on sam pling date. Fertilization w ith N did n o t affect rates o f N hxation a t any time. In addition, variation in N hxation rates m easured in unfertilized plots a t four sam pling tim e points suggested seasonal variability in N hxation: leaf litter N hxation ranged from 0.36 kg/ha/yr in the dry season to 5.48 kg/ha/yr in the wet season. Soil N hxation showed similar patterns ranging from a dry season low o f 0.26 kg/ha/yr to a wet season high o f 2.71 kg/ha/yr. W hile the observed tem poral variability suggests potential climatic control over free-living N hxation in these forests, data suggest that neither soil nor leaf litter m oisture alone regulate N hxation rates. Instead, we hypothesize that a com bination o f am ple C availability, low leaf litter N :P ratios, and high rainfall coincide during the latter portions o f the rainy season and drive the highest free-living N hxation rates o f the year. Abstract in Spanish is available a t http://www.biackweii-synergy.com /ioi/btp.in Spanish is available a t http://www.biackweii-synergy.com /ioi/btp. Key words'. C osta Rica; fertilization; free-living; phosphorus; tropical wet forest. B i o l o g i c a l n i t r o g e n (N) f i x a t i o n p r o v i d e s t h e l a r g e s t n a t ­ u r a l SOURCE OF ‘n e w ’ N TO the world’s ecosystems (GaUoway et al. 1995), but like many biogeochemical processes, N fixation rates vary widely across the terrestrial biosphere. Biome-level compar­ isons suggest that tropical rain forests may fix more N than any other unmanaged ecosystem (Stedman & Shetter 1983, Cleveland et al. 1999, Galloway et al. 2004). However, actual data assessing rates of N fixation in tropical rain forests are rare, and more recent studies suggest that N fixation by leguminous trees may not be as widespread in mature tropical rain forests as previously believed (Gebring et al. 2005). In contrast, we know less about free-living N fixation in tropical forests, and major uncertainties remain about both the rates of tropical rain forest free-living N fixation and the factors that regulate the process. If free-living N fixation rates are indeed high in tropical rain forests, it would represent an interesting paradox in ecosystem ecol­ ogy (Vitousek & Howartb 1991). Biological N fixation is energeti­ cally expensive (Gutscbick 1987), and the high cost of acquiring N via fixation should strongly disfavor the process in N-ricb ecosys­ tems, where cellular uptake should be a more efficient means of acquiring biologically available N (Vitousek et al. 2002). In ad­ dition, a high demand for adenosine tripbospbate (ATP; Sprent & Raven 1985) and phosphorus (P)-ricb cellular structures (Wolk 1982) by N-fixing organisms predisposes them to P limitation, and increased levels of P availability are known to stimulate N fixation rates in some ecosystems (Israel 1987, Pereira & Bliss 1987, Eisele Received 6 April 2006; revision accepted 27 Novem ber 2006. ^C orresponding author; e-mail: sasha.reed@ colorado.edu Present address: D epartm ent o f Ecosystems and Conservation Sciences, U niver­ sity o f M ontana, 32 C am pus Drive, Missoula, M T 59812, U.S.A. et al. 1989). The fact that many rain forests seem to maintain high N availability (Martinelli et al. 1999) and have ecosystem processes limited by P (Herbert & Fownes 1995, Hobble & Vitousek 2000, Cleveland & Townsend 2006) seems inconsistent with high rates of N fixation. On the other band, though nutritional considerations may dis­ favor free-living N fixation, the ample rainfall, warm temperatures, and large energy inputs found in tropical rain forests (Melillo et al. 1993) could promote the process. Because free-living N fixation rates are positively correlated with temperature (Hicks et al. 2003), moisture (Roskoski 1980), and labile carbon (C) availability (Vitousek & Hobble 2000), the climate and rapid C cycling typical of tropical rain forests (Raich & Nadelboffer 1989, Swift & Anderson 1989, Raich & Scblesinger 1992) could be especially conducive to free-living N fixation of the forest floor. Given the paucity of data, the high and variable rates of free-living N fixation proposed for rain forests may reflect interactions among the factors regulating N fixation rather than truly paradoxical results. Our study bad three primary goals. First, we sought to deter­ mine if free-living N fixation in leaf litter and soil was a large source of new N to a mainland tropical rain forest ecosystem. Second, we used a fiall-factorial fertilization experiment (N x P) to explore possible nutrient controls over forest floor N fixation. Third, we sought to investigate temporal patterns in leaf litter and soil N fix­ ation rates, and to explore their possible relationship to patterns in rainfall, litterfaU, and C availability We hypothesized that (1) rates of free-living N fixation would be high relative to other ecosystems; (2) fertilization with N would suppress N fixation rates, while P fertilization would enhance rates; and (3) N fixation rates would peak in the early wet season when rainfall solubilizes large stocks of labile C from the litter layer (Cleveland et al. 2006). © 2007 The Author(s) Journal compilation © 2007 by The Association for Tropical Biology and Conservation 585 586 Reed, Cleveland, and Townsend