The application of an integrated biogeochemical model to simulate dynamics of vegetation, hydrology and nutrients in soil and streamwater following a whole-tree harvest of a northern hardwood forest

Understanding the impacts of clear-cutting is critical to inform sustainable forest management associated with net primary productivity and nutrient availability over the long-term. Few studies have rigorously tested model simulations against field measurements which would provide more confidence in efforts to quantify logging impacts over the long-term. The biogeochemical model, PnET-BGC has been used to simulate forest production and stream chemistry at the Hubbard Brook Experimental Forest (HBEF), New Hampshire, USA. Previous versions of PnET-BGC could accurately simulate the longer-term biogeochemical response to harvesting, but were unable to reproduce the marked changes in stream NO₃ ⁻ immediately after clear-cutting which is an important impact of forest harvesting. Moreover, the dynamics of nutrients in major pools including mineralization and plant uptake were poorly predicted. In this study, the model was modified and parametrized allowing for a lower decomposition rate during the earlier years after the clear-cut and increased NH₄ ⁺ plant uptake with the regrowth of new vegetation to adequately reproduce hydrology, aboveground forest biomass, and soil solution and stream water chemistry in response to a whole-tree harvest of a northern hardwood forest watershed (W5) at the HBEF. Modeled soil solution and stream water chemistry successfully captured the rapid recovery of leaching nutrients to pre-cut levels within four years after the treatment. The model simulated a substantial increase in aboveground net primary productivity (NPP) from around 36% to 97% of pre-cut aboveground values within years 2 to 4 of the cut, which closely reproduced the measured values. The projected accumulation of aboveground biomass 70 years following the harvest was almost 190 t ha⁻¹, which is close to the pre-cut measured value. A first-order sensitivity analysis showed greater sensitivity of projections of the model outputs for the mature forest than the strongly aggrading forest.