A Model to Simulate the Response of a Northern Hardwood Forest Ecosystem to Changes in S Deposition

Watershed studies across the northeastern United States have shown that stream losses of SO42− exceed atmospheric sulfur (S) deposition. Understanding the processes responsible for this additional source of S is critical to quantifying ecosystem response to ongoing and potential future controls on SO2 emission. An integrated biogeochemical model, PnET-BGC, was used to investigate inputs and dynamics of S in a northern hardwood forest at the Hubbard Brook Experimental Forest (HBEF), New Hampshire, USA. The changes in soil S pools and stream-water SO42− were simulated to assess the response to both atmospheric S deposition and forest clear-cutting disturbances. Model simulation using the measured dry-to-bulk deposition ratio of 0.21 resulted in an underprediction of soil S pools and stream-water SO42− concentrations in the biogeochemical reference watershed (Watershed 6). However, the depiction of biotic processes (e.g., plant uptake, mineralization) in the model reduced the discrepancy in stream SO42− concentration between measured and model predicted value by ∼50% compared to a previous modeling effort that only considered abiotic processes. Long-term simulations (∼150 yr) indicated that elevated anthropogenic S deposition has increased stream SO42− concentrations and enhanced the incorporation of S in adsorbed SO42− and organic S soil pools. Following the implementation of the 1970 and 1990 Amendments to the Clean Air Act, decreases in S deposition resulted in the net release of S from soil pools, including soil organic S. Model simulation of forest clear-cutting of Watershed 5 at the HBEF showed that NO3− leaching and associated acidification following this disturbance increased adsorption of SO42− to soil. Compared to the reference watershed, stream-water SO42− concentrations were slightly higher, and soil adsorbed SO42− pools were substantially higher in the clear-cut watershed 4–5 yr after disturbance. Simulation of stable S isotopes showed that fractionation associated with the mineralization of soil organic S might explain the depletion in 34S observed between throughfall and stream water in the reference watershed. There is a need for further research on: (1) the rates of dry S deposition and (2) the rate of weathering of various minerals and the isotopic composition of these minerals in order to fully assess the discrepancy (i.e., greater export of S than can be accounted for by atmospheric deposition) in watershed S mass balances. The results of forecasts of the future response to anticipated decreases in S deposition are highly dependent on the nature of this additional source of S to forest watersheds. However, the large size and relatively long turnover time of soil organic S pools compared to adsorbed SO42− pools suggest that a model depicting only abiotic processes will not be suitable for predicting the long-term recovery of stream water from acidification by atmospheric deposition in northern forests.