Hurricane Effects on Nitrogen Trace Gas Emissions in Hawaiian Montane Rain Forest 1

NITROUS OXIDE (N20) IS A RELATIVELY STABLE, radiative, and active gas that is increasing in concentration in the atmosphere (Khalil & Rasmussen 1992) and is involved in reactions that lead to ozone destruction in the stratosphere (Cicerone 1987), while nitric oxide (NO) is involved in tropospheric ozone production (Thompson 1992). Several soil processes produce both NzO and NO, and tropical soils in particular are considered major sources of N trace gases (Matson & Vitousek 1989, Davidson 1991, Keller & Reiners 1994, Potter et al. 1996). Changes in emissions of N trace gases from soils have been proposed as sensitive indicators of disturbance to ecosystems (Dulohery et al. 1996). Some anthropogenic disturbances to ecosystems and changes in land use have been shown to increase NO or NzO fluxes (LuizSo et al. 1989, Davidson et al. 1991, Keller & Reiners 1994), although Dulohery et al. (1996) found no clear trend in NzO emissions following forest harvesting. The effects of natural disturbances, particularly hurricanes and tropical storms, also have been investigated. Steudler et al. (1991) examined the effect of forest damage from Hurricane Hugo on N z 0 emissions from forest soils on the island of Puerto Rico. Flux of N 2 0 from their reference stand (an area that suffered 80% defoliation but little tree mortality) was lower following Hugo (at 1, 4, 7, 11, and 14 months after Hugo) compared to one pre-Hugo value (6 mo prior). This contrasted with high emissions of N 2 0 for up to four months from an area that suffered heavy tree mortality (but which had not been sampled before Hurricane Hugo). In Massachusetts, Bowden et al. (1993) disturbed a temperate forest to mimic hurricane mortality effects. Plots in which 65 percent of the trees were damaged (uprooted, leaning or snapped) had N 2 0 emissions that were 78 percent lower in the year following treatment than the average pre-treatment emissions, and they were lower by a small amount than that of a control plot. Hurricanes occasionally have a direct impact on the forest ecosystems of Hawaii: four hurricanes made landfall in the Hawaiian islands from 1922 to 1985 (NOAA 1985). Tropical storms with winds of <74 mph are even more common, affecting Hawaii every year or two (NOAA 1985). In September 1992, Hurricane Iniki passed over a native montane rain forest on the island of Kauai in the Hawaiian islands (Fig. 1) with sustained winds of 230 kmlh and gusts up to 280 km/h (National Weather Service 1992). This forest was the oldest site in a chronosequence previously used to evaluate NzO and NO emissions from forest soils. That study began in August 1990 and ended in November 1991, less than a year before Hurricane Iniki (Riley & Vitousek 1995, Riley 1996). Two conclusions from the pre-Iniki study are important considerations when discussing hurricane effects. First, the sum of NO and NzO emissions paralleled soil N transformation rates as they varied with soil age (Riley & Vitousek 1995, Riley 1996), supporting a strong role of N availability controlling N trace gas emissions (Robertson & Tiedje 1987). Second, ratios of emissions of NzO and N O varied as a function of soil moisture, which is consistent with a conceptual model suggested by Firestone and Davidson (1989). Nitrogen availability may be influenced by hurricane disturbance through changes in litterfall and fine root biomass (Parrota & Lodge 1991, Sanford et al. 1991, Steudler et al. 1991). Soil water content may also change through hurricane effects on forest canopy interception and transpiration and soil organic matter content. Our yearlong pre-hurricane data set of NzO and N O emissions, soil moisture, and soil N transformation rates at the site affected by Hurricane Iniki offered the opportunity to directly evaluate short-term hurricane effects in this ecosystem. The site is located on Kauai in Kokee State Forest, on parent material of the Olokele member of Waimea Canyon basalt; its age is estimated at 4.1-million years (K-Ar dating; Clague & Dalrymple 1989; D. Clague, pers. comm.). The soil is mapped as a Plinthic Acrorthox (Soil Conservation Service 1972). The site is situated on a stable land surface of ca 5 percent slope, at an elevation of 1134 m. The