A linear, three-component (nitrate, phytoplankton, and zooplankton) biogeochemical model coupled to a one-dimensional turbulence closure fluid dynamic model successfully reproduces most of the thermal and biogeochemical features of three marine settings (Bermuda Station ''S'' in the oligotropic portion of the North Atlantic subtropical gyre, Ocean Weather Station ''India'' in the North Atlantic subpolar gyre, and the 60-m isobath of the New York Bight) and one freshwater lake (South Bluehill Pond in Newfoundland). Model reproduction of temperature in all four settings is good. Model temperature gradients in the thermocline are somewhat steeper than those observed in the data of the three marine settings due to the inability of the one-dimensional fluid flow model to reproduce internal waves which can be a significant source of vertical mixing in large water bodies. Nitrate and chlorophyll are predicted reasonably well by the model at Bermuda Station ''S'' and the New York Bight. Underprediction of surface nitrate at Station ''India'' is probably due to the lack of an ammonium component in the model. Model reproduction of the spring bloom productivity in the marine settings is good, although post-bloom model productivity is generally too low. Biogeochemical simulations of South Bluehill Pond are strongly dependent on the value of the light attenuation coefficient used in the model. A relatively high attenuation coefficient (most likely representing high inorganic sediment concentrations) is necessary to accurately reproduce nitrate and chlorophyll concentrations. These results illustrate the necessity of accurately characterizing the light regime prior to using the model in lacustrine settings. ¿ American Geophysical Union 1995