A physiological‐based bio‐optical model is used to estimate vertical profiles of instantaneous, diurnal, and integrated daily rates of in situ primary production throughout the water column at three stations across a coastal front. The model makes use of an empirical relationship between photosynthesis, quantum yield, and photosynthetically absorbed radiation and is a full spectral model with all relevant parameters determined as a function of wavelength. In the prior application of this bio‐optical model, parameters used to estimate quantum yield as a function of irradiance were wavelength‐independent and held constant. Here, the model is recast so that quantum yield is estimated with wavelength‐dependent photosynthesis‐irradiance parameters, Pmax the maximum rate of phytoplankton photosynthesis, and Ik the P‐I saturation parameter, which are allowed to vary with depth and time of day. Both Pmax‐dependent (assumed to be λ‐independent) and Ik‐dependent (known to be λ‐dependent) estimates of temporal/spatial changes in quantum yield were assessed. The model was tested in water masses where a net‐to‐nanoplankton transition was occurring in phytoplankton communities dominated by diatoms and prymnesiophytes. At each station there is close agreement between 14C productivity estimates derived from knowledge of diurnal patterns in P‐I parameters and productivity estimates derived from two variants of the bio‐optical model based on knowledge of spectral irradiance and phytoplankton pigmentation. The 14C and bio‐optical estimates of primary production also show similar vertical patterns for either instantaneous or daily integrated values. When compared with 14C estimates, Ik‐dependent bio‐optical estimates of photosynthetic rates give a closer match than bio‐optical estimates that are Pmax‐dependent. The model permits calculation of primary production from shipboard observations and may be useful for predicting production rates from bio‐optical data provided by untended buoys.