Primary production by marine microalgae is believed to be a critical factor regulating atmospheric carbon dioxide levels and associated climatic changes1. Assessments of photosynthesis in the open ocean2–4, and the related export of organic carbon to the deep ocean (new or non-regenerative production5–7, vary by as much as an order of magnitude (refs 2 and 6, compare with refs 3–5 and 7). Discrepancies are attributed4,6–8 to different temporal and spatial scales reflected by instantaneous rate measurements, as opposed to seasonally averaged measurements based on subsurface changes in chemical tracers. Satellite extrapolations of primary production can be used to characterize and quantify temporal and spatial variability9–11. But time differentials between satellite and ship measurements, as well as regional and seasonal variations in empirical relationships, have so far limited the precision of such extrapolations9,12. We conducted extensive ship sampling of chlorophyll a and primary production in the western Mediterranean Sea contemporaneous with Nimbus-7 coastal zone colour scanner imagery. Our approach resulted in an empirical model for estimating integrated water-column primary production from satellite imagery. Precision was adequate to resolve short-term fluctuations in primary production associated with a meso-scale circulation feature.