Ecosystem nutrient balance and dynamics

Many attempts at balancing ecosystem-level nutrient budgets have been made in the past few decades using a variety of approaches, and for a variety of different purposes. Relatively simple mass balance equations have been used at the level of the water-shed that might comprise single forested ecosystems (e.g., Binkley et al. 1982; Stohlgren et al.1991; Likens and Bormann 1995; Hedin et al. 1995; Stottlemyer and Troendle 1992), or at the level of a larger-scale region (e.g., Gold 1990; Frink 1991;Jaworski et al. 1992; Howarth et al. 1996; Valielaet al. 1997) that might include different land uses and nutrient inputs. These budgets have been developed with varying efforts at measuring, or modeling, internal processes of nutrient retention and release.
The whole-ecosystem-level budget has a variety of purposes, including:
1. directly testing mechanistic models and understanding processes that control nutrient retention, turnover, and leaching in a field setting;
2. estimating processes by difference that are difficult to measure directly (such as weathering), or to test measurement methods for complex processes when the measurement techniques are called into question (such as dry deposition inputs, nitrogen fixation, or denitrification);
3. understanding controls on regional water quality;
4. in an experimental setting, understanding responses to watershed perturbations such as harvest, fire, or anthropogenic pollutant deposition (e.g., Likens et al. 1978; Pardo et al. 1995);
5. assessing the potential for long-term depletions in critical nutrient pools, such as cation stocks in forests affected by acidic deposition (e.g.,Federer et al. 1989; Likens et al. 1996).
Each specific purpose may require a unique level of analysis and detail of measurement.
Often the lack of balance in an ecosystem budget has proven far more interesting than a budget that has balanced. For example, Bormann et al. (1977) found a higher rate of nitrogen (N) storage (as accumulation in biomass) plus loss (as streamwater) in the Hubbard Brook, NH, watershed than they found in inputs of N measured as bulk deposition. At the time, they suggested that the imbalance was probably due to unmeasured N2 fixation. Thus the imbalance led to a testable hypothesis that directed future research, and later measurements showed that fixation inputs were unlikely to correct the imbalance. Only a portion of the imbalance could be corrected by improved methods of measuring atmospheric inputs such as dry deposition. However, when the imbalance was compared with the mass of N in the soil pool (7200 kg ha-1), it became clear that a fluctuation of 0.5% in this pool would dwarf the budget imbalance, and thus much of the N uptake by vegetation probably came from a small net change in the mineral soil. This study showed the difficulty in comparing relatively well-quantified external fluxes with poorly quantified and considerably larger internal fluxes, and showed the importance of examining the common assumption of ecosystem steady state.
In this chapter I will review some of the basic methods and problems of creating whole ecosystem nutrient budgets, including budgets at the whole-watershed scale, at the smaller plot or landscape scale, and in systems without readily measurable hydrologic outflows, as well as uses of nitrogen-15 (15N) tracer techniques to follow N inecosystems. I will focus primarily on forested ecosystems, as there is relatively little data in other vegetation types at the watershed level (but see Dodds et al. (1996) for an analysis of tallgrass prairie watersheds).