Non-labile Soil 15Nitrogen Retention beneath Three Tree Species in a Tropical Plantation

T species composition affects forest biogeoSoil organic matter is the largest sink for N additions to forests. chemistry because species differ in their rates of Species composition may affect soil N retention by altering the amount nutrient and energy cycling (Zinke, 1962; Boettcher and or proportion of added N stored in non-labile organic pools. We Kalisz, 1990; Hobbie, 1992; Wardle et al., 1997; Binkley measured 15N tracer retention in labile and non-labile pools of surface and Giardina, 1998). Species–ecosystem relationships (0–20 cm) mineral soils, 7 yr after the tracer was applied to a 9 yrlink the fields of population and ecosystem ecology old Puerto Rican tree plantation with replicated stands of three species (Jones and Lawton, 1994) with major implications for (two N-fixers, one Eucalyptus, Euc). Laboratory incubations (13 mo) soil fertility, C sequestration, and plantation productivwith repeated leaching separated total soil N into labile (inorganic N ity. Species characteristics may also affect retention of leached) and non-labile (total N minus leached N) pools, and a labile recent large increases in N fertilization (Matthews, 1994; C treatment tested linkages between C availability and N retention. Binkley et al., 1995) and atmospheric N deposition (GalWe hypothesized that species composition would alter the amount and proportion of recovered tracer N in non-labile organic matter. loway et al., 1994). Surface soils contained 45% of the tracer, but the amount retained In some recent case studies, forest N retention apin labile and non-labile pools was similar among species. In contrast, pears to depend in part on species composition. In New the proportion of recovered tracer in non-labile pools was greater in England, hardwood species had higher plant 15N recovsoils beneath N-fixers (75%) than Euc (62%). Labile C additions ery (Nadelhoffer et al., 1995), total ecosystem N retetenincreased the size of the non-labile tracer N pool. We conclude that tion (Magill et al., 2000), and net N mineralization rates tree species composition may affect long-term soil N retention by (Finzi et al., 1998) than coniferous species. Many Euroaltering the proportion of N in slow-turnover, non-labile pools. Plants pean beech ( fagus stuatica L.)-dominated forests had may also alter soil N retention by renewing labile C pools; a continlower N deposition and nitrate leaching than adjacent uous supply of labile C increased the transfer of 15N into non-labile spruce (Picea A. Dietr.)-dominated forests (Rothe et organic matter. al., 2001). In contrast, a synthetic analysis of over 300 J.P. Kaye and D. Binkley, Dep. of Forest Sciences, Colorado State Abbreviations: AFE, atom fractional enrichment; ECEC, effective Univ., Fort Collins, CO 80523; X. Zou, Terrestrial Ecology Division, cation-exchange capacity; Cas, Casuarina; Euc, Eucalyptus; Leu, LeuUniv. of Puerto Rico, P.O. Box 363682, San Juan, PR 00936; and J. caena; kn, extraction efficiency; No, mass of labile N; Na, the mass of Parrotta, USDA Forest Service, Research and Development, P.O. tracer N in the labile pool; Nn, the mass of labile native soil N; No, Box 96090, Washington, DC 20090-6090. Received 22 Mar. 2001. AFE of the composite leachate sample; Na, AFE of the tracer; Nn, *Corresponding author ([email protected]). AFE of the native soil; PNDFA, percent of tree N derived from the atmosphere; NPP, net primary productivity. Published in Soil Sci. Soc. Am. J. 66:612–619 (2002). KAYE ET AL.: NON-LABILE SOIL NITROGEN-15 RETENTION BENEATH THREE TREE SPECIES 613 streams across the USA showed that conifer forests had immediately affect water quality or plant production. On the other hand, N stored in labile pools may increase lower stream-water nitrate concentrations than deciduous forests (Binkley, 2001). Among deciduous species, fluxes to plants and stream water. In agricultural systems, most fertilizer N is non labile after one growing potential net nitrification was positively correlated with atmospheric N deposition in maple (Acer L.), but not season (Broadbent and Nakashima, 1967; Stanford et al., 1970; Smith et al., 1978; Smith and Power, 1985), beech stands (Lovett and Reuth, 1999) and lysimeter nitrate concentrations were greater in east-facing stands but similar data are rare for forests (Preston and Mead, 1994b; Chang and Preston, 1998). dominated by maple and cherry (Prunus L.) than in south-facing stands dominated by gum (Nyssa sylvatica) In this paper, we report the first direct test of tree species effects on 15N-tracer retention using a replicated and beech (Peterjohn et al., 1999). In all of these studies, species effects were apparent, but were confounded by common garden experiment with three tree species (two N-fixers) and a 6 to 8 yr-old (hereafter 7 yr-old) 15N uncontrolled factors that covaried with species, such as land-use history, topography, prior soil conditions, or addition. Rather than focusing on traditional plantavailable N pools, we estimated species effects on nonregional climate. Common garden experiments are needed to discern species effects on N loss and retention labile pools of soil organic N. We hypothesized that tree species would affect the amount and proportion of in plots with similar soils, topography, and climate (e.g., Johnson and Todd, 1988). added N that was retained in non-labile soil pools. To determine whether species effects resulted from differTree species could affect N retention by several mechanisms, the most obvious being differential plant uptake ences in labile C inputs, we also conducted a labile C addition experiment. of N. While this mechanism may be important, most 15N-tracer experiments show that soils, rather than There is no standard method to separate organic N into labile and non-labile pools. Previous studies have plants, are the largest sink for N added to forests (Mead and Pritchett, 1975; Heilman et al., 1982; Melin et al., used physical (Strickland et al., 1992) and chemical (He et al., 1988) methods to fractionate soil C and N, as1983; Clinton and Mead, 1993; Preston and Mead, 1994a; Tietema et al., 1998; Nadelhoffer et al., 1999). Tree suming that aggregate size, organic matter density, or organic matter solubility was well correlated with C species could alter soil N retention if the quantity or quality of C inputs differs among species. Differences or N availability to microorganisms. Biological fractionations, such as the long-term incubations used in this in C inputs could alter microbial immobilization of N (Jansson, 1958; Kelley and Stevenson, 1987; Hart et al., study, allow the “in situ microbial and microarthropod community to define ecologically relevant SOM (soil 1994), and differences in litter quality may affect the amount of N sequestered in non-labile humus (Melillo organic matter) fractions” (Robertson and Paul, 1999). et al., 1989; Berg and Matzner, 1997). Large N inputs MATERIALS AND METHODS from N-fixing tree species may saturate soil N sinks and limit further N retention (Van Miegroet et al., 1990; The study site is on the northern coast of Puerto Rico at the Binkley et al., 1992), but N fixers may also promote University of Puerto Rico’s Toa Baja Agriculture Experiment Station (Parrotta et al., 1993, 1996; Parrotta, 1999). The plantanon-labile organic matter formation (Kaye et al., 2000; tion was organized as a completely randomized block (n 3 Resh et al., 2001) and thus, long-term soil N sequesblocks) experiment with six treatments per block. The treattration. ments originally applied were monocultures and mixtures of Most analyses of species effects on N retention have Eucalyptus robusta J.E. Smith (Euc), N-fixing Casuarina equifocused on changes in actively cycling, labile N pools setifolia J.R. & G. Forst. (Cas), and N-fixing Leucaena leuco(inorganic or microbial biomass N). However, most soil cephala (Lam.) de Wit (Leu). For research described here, N is not actively cycled by plants and microbes on annual we sampled only the monocultures at 9 yr of age. The stands time scales, and this large non-labile N pool may be an were planted in 16 by 16 m plots at a spacing of 1 by 1 m. important sink for N additions. Nitrogen retained in Some general characteristics of the plantations are presented in Table 1. a non-labile pool with a slow turnover time will not Table 1. Some characteristics of biomass and soils in the plantations. From Parrotta (1999) and Parrotta et al. (1996).