113-119 Matson.p65

Human activities have more than doubled inputs of nitrogen (N) to terrestrial ecosystems worldwide (1). Until recently, most of this increase was focused in the developed regions of the world. With fertilizer consumption (2), fossil-fuel emissions (3), and biomass burning (4) rapidly increasing in less developed regions, sources and distribution of anthropogenic N now occur globally (5), as does the deposition of N. Anthropogenic N additions to temperate ecosystems have been shown to affect a wide range of ecosystem properties and processes, especially when inputs are large and continuous. For example, N additions can alter net primary production (NPP) and nutrient cycling (6), interact with elevated carbon dioxide (CO2) (e.g. 7), cause significant reductions in biological diversity (8), and influence rates of N transfer via hydrologic loss (9). As discussed in Rabalais (10), increased nitrate leaching from N-saturated ecosystems also drives acidification and eutrophication and alters biodiversity in downstream freshwater and marine systems (e.g. 11). Many of the observed responses of terrestrial systems to N deposition can be understood within the context of “nitrogen saturation” (12, 13), a conceptual model that was developed using knowledge derived from temperate-forest ecosystems and that has been applied to changes in ecosystem functioning that occur as N additions relieve N limitation to biological activity. In general, the N saturation model predicts that N-limited systems initially retain anthropogenic N by using it for plant and microbial growth as well as accumulation in biomass and soil organic matter. At some point, however, inputs of N begin to exceed the biotic (and perhaps abiotic) demands for N within the ecosystem, and the system is predicted to lose its N retention capacity. As the capacity to retain N is exceeded, excess N is available to be lost from the system via solution losses and gas fluxes. For example, nitrate leaching has been identified as a primary sign of N saturation in N-limited temperate forests if it occurs at a significantly higher rate than “background” levels, immediately after N additions without a significant temporal lag period, or aseasonally in a naturally seasonal ecosystem (14). Furthermore, the model predicts eventual decreases in production that can occur due to cation losses, nutrient imbalances, or increased plant susceptibility to other stresses, such as frost, ozone, or insect attack. Not all terrestrial ecosystems, however, respond to elevated N inputs in these ways. In recent years, the accumulation of information about the effects of N deposition in a variety of ecosystems, primarily in North America and Europe, has allowed analysis of the range of responses among terrestrial ecosystems, as well as the factors that contribute to that dramatic range (e.g. 15). In addition, the globalization of N deposition raises questions concerning consequences of anthropogenic N for ecosystems outside the well-studied regions. Because tropical forests cover 17% of the Earth’s surface and make up 40% of the global NPP (16), even small alterations in major biogeochemical cycles there can have global significance. Relatively young soils common to temperate and boreal ecosystems differ markedly from highly-weathered tropical systems; while N generally limits productivity in temperate ecosystems (17), phosphorus (P) or calcium (Ca), removed from soils through intense weathering processes most likely limit production in humid tropical forest and savanna ecosystems (18). In many of these systems, N may already be a nutrient in excess, so many tropical humid forests may be naturally N saturated (19). In this paper, we summarize some of the direct and indirect effects of increasing anthropogenic N inputs on ecosystem processes in temperate ecosystems and compare some of these effects to predicted responses in tropical ecosystems. We draw on a number of excellent reviews on the issue of N deposition and saturation in terrestrial ecosystems that have been published in the past several years (for example, 6, 8, 15, 20–22). Thus, we do not attempt a thorough review, but rather we point to some of the conditions under which anthropogenic impacts can be significant, some of the factors that contribute to variations in response, and some of the areas where uncertainty is large due to limited information.