A simple ecosystem model is coupled to a 3-dimensional general circulation model for the North Atlantic. The physical model is based on the Los Alamos Parallel Ocean Program (POP) and forced by climatological monthly mean data. Four biological components(phytoplankton, zooplankton, nutrients and detritus) are incorporated into POP as additional tracers with biological sources and sinks. The model solutions, obtained with different physical and biological parameterizations are compared against monthly mean SeaWiFS colour data averaged over the period 1997-2003 and Levitus's climatological nitrate data. A reference model solution, with constant biological model parameters over the whole basin,underestimates both the average chlorophyll level and its regionalvariability at mid- to high-latitudes. Experiments with a different parameterization of heat and freshwater fluxes, which affects upper ocean mixing, indicate a strong impact of such parameterizations on nutrient supply to the surface layer at high latitudes, but with little impact on simulated chlorophyll. Other experiments where advection of the biological tracers is turned off show basically the same result: strong impact on regional nutrient patterns but a negligible impact on phytoplankton patterns. Only model runs with spatially variable biological parameters, obtained from a previouszero-dimensional ecosystem model calibration on CZCS ocean colour data, could reproduce regional scale patterns in the SeaWiFS imagery. We hypothesize that some of these patterns can be linked to coccolithophore blooms in areas influenced by the N. Atlantic Drift during summer and to effects of temperature on plankton loss ratesduring spring. Future work should focus on identifying the main factors responsible for these spatial patterns and developing the ecosystem models that can capture them.