Abstract. The analysis of measured nutrient concentrations suggests that the ratio of dissolved inorganic nitrogen (DIN) alteration before and after spring blooms relative to the alteration of dissolved inorganic phosphorus (DIP) remains quite constant over the years (2000~2009). This ratio differs from the Redfield ratio and varies from 6.6 : 1 to 41.5 : 1 across basins within the Baltic Sea. If the found N/P ratios are indicators of phytoplankton stoichiometry, this would affect nutrient cycles in ecosystem models. We therefore tested the effects of using horizontally variable N/P ratio instead of fixed ratio (N/P = 10 : 1 or 16 : 1) on phytoplankton uptake and remineralization in a 3-D physical-biogeochemical coupled model ERGOM. The model results using the variable N/P ratio show systematical improvements in model performance in comparison with the fixed ratios. In addition, variable N/P ratios greatly affected the model estimated primary production, nitrogen fixation and nutrient limitation, which highlights the importance of using an accurate N/P ratio.