Nitrogen leaching from N limited forest ecosystems: the MERLIN model applied to Gårdsjön, Sweden
Abstract. Chronic deposition of inorganic nitrogen (N) compounds from the atmosphere to forested ecosystems can alter the status of a forest ecosystem from N-limited towards N-rich, which may cause, among other things, increased leaching of inorganic N below the rooting zone. To assess the time aspects of excess N leaching, a process-oriented dynamic model, MERLIN (Model of Ecosystem Retention and Loss of Inorganic Nitrogen), was tested on an N-manipulated catchment at Gårdsjön, Sweden (NITREX project). Naturally generated mature Norway spruce dominates the catchment with Scots pine in drier areas. Since 1991, ammonium nitrate (NH4NO3) solution at a rate of about 35 kg N ha-1 yr-1 (250 mmol m-2 yr-1) has been sprinkled weekly, to simulate increased atmospheric N deposition.
MERLIN describes C and N cycles, where rates of uptake and cycling between pools are governed by the C/N ratios of plant and soil pools. The model is calibrated through a hindcast period and then used to predict future trends. A major source of uncertainty in model calibration and prediction is the paucity of good historical information on the specific site and stand history over the hindcast period 1930 to 1990. The model is constrained poorly in an N-limited system. The final calibration, therefore, made use of the results from the 6-year N addition experiment. No independent data set was available to provide a test for the model calibration.
The model suggests that most N deposition goes to the labile (LOM) and refractory (ROM) organic matter pools. Significant leaching is predicted to start as the C/N ratio in LOM is reduced from the 1990 value of 35 to <28. At ambient deposition levels, the system is capable of retaining virtually all incoming N over the next 50 years. Increased decomposition rates, however, could simulate nitrate leaching losses. The rate and capacity of N assimilation as well as the change in carbon dynamics are keys to ecosystem changes. Because the knowledge of these parameters is currently inadequate, the model has a limited ability to predict N leaching from currently N-limited coniferous forest ecosystems in Scandinavia. The model is a useful tool for bookkeeping of N pools and fluxes, and it is an important contribution to further development of qualitative understanding of forest N cycles.