MERLIN model applied to NITREX Gårdsjön

Abstract

A new process-oriented dynamic model, MERLIN, has been tested on a nitroen-manipulated catchment at Gårdsjön, Sweden, to elucidate the time aspects of nitrogen "saturation". The model is calibrated to the period 1930-90 and the first 4 years of N addition (1991-94), and then compared with data for the 5th and 6th years. The calibrated model is used to predict future trends in N leaching. The model indicates that most of teh added N goes to forest floor soil and very little to the plants. This agrees with measured N concentrations in needles. The model calibrated well to the flux of N out of the catchment during the first 4 treatment years, and correctly predicted the 5th year but overestimated the 6th year. In the model, significant leaching occurs as the optimum C/N ratio in forest floor is reduced from the 1990 value of about 35 to about 25. This agrees with the general results from the European NITREX sites, at which significant N leaching occurs at sites with low C/N ratio in the forest floor. The time needed for a system to change from N-limited, indicated by a high C/N, to a N-rich system, with C/N ratio below 25 and increased N leaching, is, however, uncertain, and the processes that govern such a transition are not fully understodd. A sensivity analysis is needed with further testing on a variety of forest ecosystems evaluate the utility of the model.