71 Ecological Applications, 15(1), 2005, pp. 71–86  2005 by the Ecological Society of America TERRESTRIAL C SEQUESTRATION AT ELEVATED CO 2 AND TEMPERATURE: THE ROLE OF DISSOLVED ORGANIC N LOSS EDWARD B. RASTETTER, 1,4 STEVEN S. PERAKIS, 2 GAIUS R. SHAVER, 1 AND GO ¨ RAN I. A ˚ GREN 3 1 The Ecosystems Center, Marine Biological Laboratory, Woods Hole, Massachusetts 02543 USA 2 U.S. Geological Survey, Forest and Rangeland Ecosystem Science Center, Corvallis, Oregon 97331 USA 3 Department of Ecology and Environmental Research, Swedish University of Agricultural Sciences, Box 7072, SE-750 07 Uppsala, Sweden Abstract. We used a simple model of carbon–nitrogen (C–N) interactions in terrestrial ecosystems to examine the responses to elevated CO 2 and to elevated CO 2 plus warming in ecosystems that had the same total nitrogen loss but that differed in the ratio of dissolved organic nitrogen (DON) to dissolved inorganic nitrogen (DIN) loss. We postulate that DIN losses can be curtailed by higher N demand in response to elevated CO 2 , but that DON losses cannot. We also examined simulations in which DON losses were held constant, were proportional to the amount of soil organic matter, were proportional to the soil C:N ratio, or were proportional to the rate of decomposition. We found that the mode of N loss made little difference to the short-term (60 years) rate of carbon sequestration by the ecosystem, but high DON losses resulted in much lower carbon sequestration in the long term than did low DON losses. In the short term, C sequestration was fueled by an internal redistribution of N from soils to vegetation and by increases in the C:N ratio of soils and vegetation. This sequestration was about three times larger with elevated CO 2 and warming than with elevated CO 2 alone. After year 60, C sequestration was fueled by a net accu- mulation of N in the ecosystem, and the rate of sequestration was about the same with elevated CO 2 and warming as with elevated CO 2 alone. With high DON losses, the ecosystem either sequestered C slowly after year 60 (when DON losses were constant or proportional to soil organic matter) or lost C (when DON losses were proportional to the soil C:N ratio or to decomposition). We conclude that changes in long-term C sequestration depend not only on the magnitude of N losses, but also on the form of those losses. Key words: carbon–nitrogen interactions; carbon sequestration; dissolved inorganic nitrogen; dissolved organic nitrogen; ecosystem models; global climate change; soil C:N; terrestrial ecosystems. INTRODUCTION Terrestrial ecosystems are thought to sequester 25% of the carbon (C) currently emitted through fos- sil-fuel burning and land use change (IPCC 2001). It is hoped that these ecosystems will continue to be a major sink for C in the future and thereby mitigate further increases in CO 2 in the atmosphere. However, productivity in terrestrial ecosystems is strongly con- strained by the dynamics of the nitrogen (N) cycle (Vi- tousek et al. 1998) and C sequestration is likely to require a net accumulation of N in these ecosystems. The input of N to ecosystems has been widely studied, especially from the perspective of atmospheric N de- position (Ollinger et al. 1993, Galloway et al. 2003, 1995) and an understanding of the controls on biolog- ical N 2 fixation is emerging (Cleveland et al. 1999, Rastetter et al. 2001, Vitousek et al. 2002). However, surprisingly little is known about the form, magnitude, or controls of N losses from terrestrial ecosystems (Sol- lins and McCorrison 1981, Hedin et al. 1995, Aber et Manuscript received 6 October 2003; revised 17 May 2004; accepted 7 June 2004. Corresponding Editor: A. R. Townsend. 4 E-mail: erastett@mbl.edu al. 2002, McDowell 2003, Neff et al. 2003; Pellerin et al., 2004). In this paper we argue that the amount of C sequestered in terrestrial ecosystems in response to elevated CO 2 depends on the fraction of N losses that are in the form of dissolved organic N (DON) vs. dis- solved inorganic N (DIN). Because plants can curtail DIN losses as N demand increases in response to el- evated CO 2 , but plants have little control over DON losses, the potential for accumulating N by limiting N losses should be small if DON losses are high. Thus, the potential for sequestering C in response to elevated CO 2 should be small if a large fraction of the N losses are as DON. Modifications to the Standard Model Our assessment of C sequestration in relation to DON losses relies upon three modifications to what has been called ‘‘the standard model’’ of N accumulation in terrestrial ecosystems (Vitousek et al. 1998). First, as suggested by Vitousek et al. (1998) and Neff et al. (2003), the standard model needs to be modified to include DON losses. Second, the standard model needs to be modified to accommodate an increase in N de- mand by both plants and microbes in response to el-