Modeling leaching as a decomposition process in humid montane forests

Ecology, Sept, 1997 by William S. Currie, John D. Aber

INTRODUCTION

Linked models of production and decomposition attempt to characterize the dominant controls on C and nutrient stores and interactions in ecosystems (Pastor and Post 1986, Rastetter et al. 1991). A process model coupled to a GIS (Geographic Information System) enables investigators to model landscape-scale patterns in ecosystem function (Running and Coughlan 1988, Burke et al. 1990, King 1991, Aber et al. 1993a). The success of such an approach depends upon capturing the dominant controls on the processes of interest across patches and gradients present within the region of interest (Gosz 1992, Alperin et al. 1995).

At the foundation of virtually all decomposition models is a single or multiple empirical constant, k in the expression [e.sup.-kt], indicating the rate of detrital mass loss over time (Olson 1963). Leaching and C[O.sub.2] mineralization are two separate mechanisms of mass loss (Gosz et al. 1973, Berg and Staaf 1980b, Schlesinger and Hasey 1981, McClaugherty 1983, Cronan 1985, Yavitt and Fahey 1986). However, because of the difficulty of distinguishing between these mechanisms in the field, mass loss in litter decay studies has typically been measured as a single process (e.g., Abet et al. 1984, Fahey et al. 1988).

Fluxes of the downward leaching of organics from the forest floor (the O horizon) into mineral soil have been made in a few humid, temperate forests (McDowell and Likens 1988, Qualls et al. 1991, Vance and David 1991, Currie et al. 1996). Compounds that comprise dissolved organic carbon (DOC) in forest-floor leachate have been characterized as primarily high-molecular-mass, complex organic acids (Cole et al. 1984, David et al. 1989, Vance and David 1989, Quails and Haines 1991, 1992, Cronan et al. 1992). These can be categorized as humic substances or prehumic substances, consisting of incomplete products of decomposition and microbial modification (Cronan and Aiken 1985, Guggenberger and Zech 1994).

Our goal was to synthesize current understanding of qualities and quantities of leachate from forest floors in a process model of decomposition that included leaching, and to predict patterns in leaching relative to other processes in a heterogeneous landscape. We studied relationships among types of forest, rates of litter input, fluxes of C[O.sub.2] mineralization, mass of the forest floor, capital of organic N, and leaching of DOC and DON (dissolved organic nitrogen) as they varied with elevation in an upland-montane region. The decomposition model derives from a series of field studies and analyses of litter decomposition and N dynamics conducted in temperate forests over the past 15 yr (Aber and Melillo 1982, Melillo et al. 1982, McClaugherty et al. 1985, Melillo et al. 1989, Aber et al. 1990). Major pools in the model represent carbon fractions, operationally defined as extractives, acid-soluble material (holocellulose) and acid-insoluble material (conventionally referred to as "lignin") (Berg et al. 1984, Taylor and Parkinson 1988, White et al. 1988, Harmon et al. 1990, Ryan et al. 1990, Means et al. 1992). We included microbial turnover in order to account for the production of secondary compounds (Van Veen et al. 1984). We included N dynamics as transfers of N among the C fractions and thereby accounted for measured fluxes of dissolved organic N (DON) in leachate from the forest floor. Because it simulates leaching fluxes, we call the model DocMod (Dissolved Organic Carbon Model).

The ability to model DOC and DON fluxes resulting from incomplete decomposition processes in the forest floor distinguishes DocMod from other decomposition models. DocMod is currently being used in a standardized model intercomparison as part of the LIDET study (Long-Term Ecological Research Network Intersite Decomposition Experiment Team), a broad field experiment designed to test the effects of substrate quality and macroclimate on decomposition processes (M. E. Harmon, unpublished manuscript; D. Moorhead, unpublished manuscript). The other models included in the comparison, CENTURY (Parton et al. 1987), GENDEC (Moorhead and Reynolds 1991), and MBL-GEM (Rastetter et al. 1991) refer to C fractions as determinants of C and N dynamics in detritus, but do not distinguish leaching as a mechanism of detrital mass loss. CENTURY, for example, was originally developed for grassland ecosystems. Leaching is typically not an important mechanism of transfer of C and nutrients in semiarid grasslands (Burke and Lauenroth, in press).

In humid forests, organics eluviated downward from the forest floor comprise important transfers of C and nutrients to mineral soil (Schoenau and Bettany 1987, Qualls et al. 1991). When DON is measured, it is typically found to be the dominant form of N leaching from the forest floor in undisturbed forests (Sollins and McCorison 1981, Fahey et al. 1985, Quails et al. 1991, Currie et al. 1996). DON was found to be the dominant form of N exported in streams from an unpolluted, old-growth, temperate forest in Chile (Hedin et al. 1995). Production of DON in the forest floor, followed by uptake of DON, has also recently been implicated as a means of competition for N between mycorrhizae (and their plant hosts) and free-living microorganisms in soil (Northup et al. 1995). The synthesis presented here should help to bridge the gap between studies of decomposition and studies of the leaching and transport of dissolved organics.