We measured metabolism rates in a shallow, nitrogen-enriched coastal marine ecosystem on Cape Cod (MA, USA) during seven summers using an open-water diel oxygen method. We compared two basins, one directly receiving most of the nitrogen (N) load (“Snug Harbor”) and another further removed from the N load and better flushed (“Outer Harbor”). Both dissolved oxygen and pH varied greatly over the day, increasing in daylight and decreasing at night. The more N-enriched basin frequently went hypoxic during the night, and the pH in both basins was low (compared to standard seawater) when the oxygen levels were low, due to elevated carbon dioxide. Day-to-day variation in gross primary production (GPP) was high and linked in part to variation in light. Whole-ecosystem respiration tended to track this short-term variation in GPP, suggesting that respiration by the primary producers often dominated whole-system respiration. GPP was higher in the more N-loaded Snug Harbor. Seagrasses covered over 60 % of the area of the better-flushed, Outer Harbor throughout our study and were the major contributors to GPP there. Seagrasses covered 20 % of the area in Snug Harbor for the first 5 years of our study, and their contribution to GPP was relatively small. The seagrasses in Snug Harbor died off completely in the 6th year, but GPP remained high then and in the subsequent year. Overall, rates of phytoplankton GPP were relatively low, suggesting that benthic micro- and macro-algae may be the dominant primary producers in Snug Harbor in most years. Net ecosystem production in both Snug Harbor and the Outer Harbor was variable from year to year, showing net heterotrophy in some years and net autotrophy in others, with a trend towards increasing autotrophy over the 7 years reported here.