Depth-integrated primary production (ΣP, in grams of carbon per square meter per day) was measured using 14 C in the northern San Francisco Estuary (SFE) from March through August of 2006 and 2007. Determinations of ΣP were then used to calibrate a published light-utilization model that relates ΣP to a composite parameter of chlorophyll, solar irradiance, and photic zone depth. The resultant calibration coefficient, y, varied by a factor of nearly two between 2006 and 2007 and was lower than determined in previous calibrations for the estuary. The now chronically low chlorophyll concentrations in the SFE have resulted in lower predictive power of the lightutilization model. The variation in y was likely the result of interannual variation in phytoplankton assimilation number. These results suggest that using a single y may yield large errors in estimated estuarine production when applied overbroad spatial and temporal scales. Given the food-limited condition of the SFE, it appears that direct measurements of primary production are necessary for accurately characterizing the base of the estuarine food web.