Global nitrogen dynamics have changed dramatically since the turn of the century. Each year an increasing amount of atmospheric nitrogen is fixed through combustion, fertilizer production and legume cultivation. These increases may have significant effects on terrestrial and aquatic ecosystems yet these effects are not well understood. To address this question, a Scientific Committee on Problems in the Environment (SCOPE) led by Bob Howarth of Cornell held a meeting in Block Island, Rhode Island. The result of this meeting was an in-depth examination of the nitrogen dynamics of the major watersheds that empty into the North Atlantic Basin.

The highest rates of global fertilizer applications are found in Europe, the Eastern U.S. and China. The Northern Atlantic region was chosen for this study because it is bounded by both Europe and the Eastern U.S. To examine the effects of increased nitrogen input to the regions that surround the North Atlantic, the region was divided into 14 watershed regions that spanned a wide range of social and environmental characteristics such as population density, fertilizer use, N deposition, soil type, and vegetation composition. At first glance, nitrogen efflux in watersheds was well correlated with population density, with some distinct and interesting differences between the tropics and temperate latitudes.

One of the primary questions of the SCOPE study was to determine the degree to which nitrogen fluxes have changed in the watershed regions since pre-industrial times. To do this, major anthropogenic nitrogen inputs and outputs were estimated for each watershed region. Ammonia was not included in deposition inputs because it comes largely from fertilizer and would be counted twice if included again in deposition. Fertilizer and industrially derived deposition (NO_y) are the principle inputs to most of the regions. Figure 22.1 shows the correlation between fertilizer application and riverine N flux.

In order to examine the impacts of anthropogenic activity on the nitrogen cycle in the watersheds, total watershed N inputs were compared to riverine N flux. This comparison resulted in a roughly linear correlation across more than an order of magnitude of variation in both values. Simple regressions of fertilizer or NO_y deposition versus river N also showed significant linear correlations in each case, and surprisingly, the correlation with deposition was stronger.

It is difficult to determine the fluxes of nitrogen prior to the influence of humans as it is usually necessary to substitute space for time by examining fluxes from pristine regions. Estimates for baseline fluxes from the temperate regions in this study ranged from 74 to 114 kg N km ^-2 yr^-1 (see Table 22.1). These values represent a number of independent data sources ranging from Canada to Oregon to 1890s Connecticut River data. When current fluxes from the North Atlantic watersheds are compared to these baseline estimates, the ratios of current to baseline fluxes range from 3.4 to 5.3 in North America and 6.8 to 10.6 in Europe (see Table 22.2). Thus, even if there are errors in estimates in the baseline flux estimates, it is clear that efflux from many of the Northern Atlantic regions has been significantly increased due to anthropogenic forces.

Despite the increases in river N efflux in many of the North Atlantic watersheds, most of the excess N is not reaching the rivers. In general, the ratios of river N exports to N inputs in temperate regions were low and ranged from 20 to 40%. This result indicates that 60 -80% of the nitrogen going into terrestrial ecosystems is being retained or transformed to N₂. The fate of this excess nitrogen is not currently known, but the resolution of this question is critical to our understanding of nitrogen cycling, as many of the possible upstream pathways may eventually saturate, thereby increasing N loads to aquatic systems and the coastal ocean.

Interestingly, the ratio of river N efflux to N inputs in the Amazon/Tocantins basin is 2.5, and in the Central America/Orinoco basin is 0.88. This result highlights one of the major differences between temperate and tropical latitudes. Whereas N limitation of plant growth in temperate ecosystems appears nearly ubiquitous, humid tropical ecosystems are relatively N rich and are not likely to tightly retain excess N inputs. Because the most dramatic increases in N deposition and fertilizer use in the next century will be in the tropics, the probability of increased losses to both atmospheric and aquatic environments appears quite high.

One of the major outstanding questions from this study is why river N effluxes are not increasing with the same exponential trend that inputs exhibit. This raises the important question of when terrestrial ecosystems will become N saturated. The nitrogen that is missing in the current mass balance estimates may experience several different fates. One possibility is that increased nitrogen deposition increases carbon uptake in vegetation and thus results in nitrogen and carbon storage in biomass. However carbon and nitrogen storage due to nitrogen deposition is not likely to occur in agricultural systems where there is ample N due to fertilizer use. This leaves only the natural systems to provide such a sink for excess N. At best, 24% of all anthropogenic inputs could reach natural systems, thus, even with 100% retention of this N, natural systems can only account for one forth of the missing N. The other possibilities for N storage are aquifer storage of nitrate and denitrification to N₂. Using nitrate increases in ground water (with poor data), it still appears that all the N cannot be accounted for. Some combination of these three factors is likely to be responsible for the retention of N in terrestrial ecosystems.

Several conclusions were drawn from this study. One, only 20-30% of anthropogenic inputs of N to temperate watersheds could be accounted for in the rivers. Two, river N efflux, from most areas, still appears to be several fold higher than in pristine times. Three, both agricultural (fertilizer use and legume cultivation) and industrial activity (NO_y deposition) appear to influence river N values. Finally, tropical regions still appear relatively unchanged, but the projected increases in N loading combined with a naturally rich N cycle suggests that N losses from tropical regions may increase dramatically in the coming decades.