A one-dimensional ecosystem model applied to time-series Station KNOT

A vertically one-dimensional ecosystem model is applied to Station KNOT (Kyodo North Pacific Ocean Time series; 44degreesN, 155degreesE). This model has 15 compartments, including two categories of phytoplankton (diatoms and non-diatom small phytoplankton) and three categories of zooplankton (small, large and predatory zooplankton). Observed seasonal cycles of ecosystem dynamics at Station KNOT, such as surface nutrient concentrations and column-integrated chlorophyll-a, are successfully reproduced by the model. Observed significant seasonality of a total primary production is also reproduced, but its amount is overestimated by more than 50%, especially during summer. The simulated spring diatom bloom seems to occur I month earlier than in reality, considering dramatic decreases in the observed surface silicate in May and June, but it is impossible to determine the time of the actual bloom more precisely because of lack of data in April. Sensitivity studies for several important parameters are described. The Ekman upwelling velocity strongly affects the amount of the total primary production. Stoichiometry of silicon to nitrogen for diatoms strongly determines the amount of the primary production by small phytoplankton. Maximum photosynthetic rate for diatoms contributes to set both the timing and strength of the spring diatom bloom. Maximum rate of grazing on diatoms by large zooplankton controls the timing of the end of the spring diatom bloom and the strength of the autumn diatom bloom. While most of the parameters can be set to the same values as those for Station A7 (41.5degreesN, 145.5degreesE), in the Northwestern Pacific like Station KNOT, some values need to be modified. The modification of the parameter values may be partly caused by the difference in the ecosystem dynamics between Stations KNOT and AT Observations of the mixed-layer depth and surface nutrient concentrations in March and April, when they reach their maxima, are strongly required to reduce uncertainty of the parameter values. (C) 2002 Elsevier Science Ltd. All rights reserved.