Nutrient dynamics and nitrogen trace gas flux during ecosystem development in montane rain forest

Ecology, Jan, 1995 by Ralph H. Riley, Peter M. Vitousek

INTRODUCTION

Soil age is a major factor underlying variation in soils and ecosystems (Jenny 1961). One of the important changes that occurs with increasing soil age is an increase in the size of nitrogen (N) pools and the rate of N cycling. Nitrogen is virtually absent from most parent materials at the beginning of soil development (Walker and Syers 1976, Vitousek and Walker 1989). Subsequently, N fixation and deposition lead to an accumulation of N in terrestrial ecosystems (Dickson and Crocker 1953, Walker and Syers 1976, Vitousek et al. 1983, Birkeland 1984, Robertson and Tiedje 1984, Vitousek et al. 1989, Bormann and Sidle 1990). The biological availability of N increases concurrently (Robertson and Vitousek 1981, Vitousek et al. 1983, Robertson and Tiedje 1984, Vitousek et al. 1989). Initially, chronosequences often correspond to gradients of increasing soil N transformation rates and N availability.

Emissions from soils of the N trace gases nitrous oxide ([N.sub.2]O) and nitric oxide (NO) can reflect patterns of N cycling (Matson and Vitousek 1987, Davidson 1991), and these patterns may therefore be useful in predicting emissions of the gases regionally and globally (Matson and Vitousek 1990). [N.sub.2]O is a long-lived, radiatively active gas and a precursor of stratospheric NO, which catalyzes ozone destruction in the stratosphere (Cicerone 1987); its concentrations are increasing globally (Robertson 1993). Nitric oxide is a highly reactive gas in the troposphere where it is a critical determinant of the concentration of the tropospheric ozone (Thompson 1992).

Firestone and Davidson (1989) suggested a conceptual model of NO and [N.sub.2]O emissions they termed the "hole-in-the-pipe" model. This model uses the analogy of a leaky pipe to suggest that there are three levels of controls that regulate the emissions of NO and [N.sub.2]O from the soil to the atmosphere: (1) factors controlling the rates of denitrification and nitrification (the "flow through the pipe"); (2) factors regulating the proportions of end products of these processes (the "size of the holes in the pipe"), and (3) factors controlling the consumption of these gases within the soil matrix.

At the first level of regulation, emissions of the trace gases NO and [N.sub.2]O may be considered to be proportional to soil N transformation rates. For example, Matson and Vitousek (1987) showed that [N.sub.2]O fluxes from humid tropical forest soils were correlated with indices of net N mineralization. In a chronosequence in which N is accumulating and N transformation rates increase with soil age, emissions of N trace gases would be expected to increase as well.

A second point of control on the production of NO and [N.sub.2]O is on the ratio of possible end products of nitrification or denitrification. Denitrification can produce NO, [N.sub.2]O, and [N.sub.2]; the proportions of these gases vary as a function of several factors, including acidity and the availability of nitrate and organic carbon to denitrifiers (Firestone et al. 1979, Davidson 1991). Nitrous oxide production is favored over [N.sub.2] production by denitrifiers when [N[O.sub.3].sup.-] levels are high relative to available carbon. Thus, if a chronosequence represents a transition from low N availability early in succession to higher N availability later in succession, [N.sub.2]O and perhaps NO may increase as a proportion of total denitrification gas flux.

The third level of control emphasizes factors that affect the probability that gases produced within the soil will reach the soil surface and atmosphere. Soil texture, structure, and water content influence the diffusivity of the N gases; high soil water content or finer soil texture increase the path length for diffusion of [N.sub.2]O and NO from the site of production to the soil surface and increases the opportunities for further reduction of these gases by denitrifiers.

In this paper; we report patterns of N mineralization and nitrification and N trace gas emission rates in five sites arrayed along a very long (200-4.5 x [10.sup.6] yr) soil age gradient in the Hawaiian islands. We measured N pool sizes (in both soil and foliage), soil N mineralization and nitrification rates (both net and gross), and field emissions of both [N.sub.2]O and NO for a 1-yr period. We anticipated that the magnitude and composition of N trace gas emissions at sites along the 4.5 x [10.sup.6] yr chronosequence would reflect the expected increase in N availability with increasing soil age. We further expected that this unusually long chronosequence would be useful in developing insights into general mechanisms controlling emissions of these gases.

STUDY SITES

In Hawaii, the age of parent material generally increases as one moves from the southeast to the northwest end of the island chain, a result of the north-westward passage of the Pacific Plate over the Hawaiian hotspot (Clague and Dalrymple 1989). Kilauea volcano on the southeast edge of the Island of Hawaii is active, while Kauai on the northwest end of the high islands is [approximately equal to]4.5 x [10.sup.6] yr old. The material erupted from Hawaiian volcanoes is almost exclusively classified as basalt or associated differentiated lavas (Langenheim and Clague 1987). Four major eruptive stages of Hawaiian volcanoes are recognized, with each stage defined in part by differences in the chemical composition of the magma produced. Variation in the content of basic cations helps distinguish lavas of the two eruptive stages upon which these sites are located. The postshield volcanic lavas and tephra underlying the three older sites are considered more alkalic (higher content of basic cations) than the shield-stage lavas and tephra (Clague and Dalrymple 1989) upon which the two youngest sites have developed. The influence of these differences on ecosystem development in this chronosequence is uncertain, though we consider the time separating the young sites (200 and 6000 yr old) from the older sites (185 000 to 4.5 x [10.sup.6] yr old) to be the dominant influence on ecosystem properties. Igneous rock, including basalts, generally contains extremely small amounts of N (Stevenson 1962, Letolle 1980). In this study, plant-available nitrogen is considered to be derived almost entirely through biological activity and atmospheric deposition.