The present day contemporary CO2 fluxes in shelf seas could be significant for the global carbon cycle, since available estimates converge to a sink of similar to 0.3 PgC yr(-1) corresponding to 21% of most recent estimate of contemporary sink of atmospheric CO2 in open oceans of 1.4 PgC yr(-1). These estimates are prone to large uncertainty mainly due to inadequate representation of the spatial variability and need to be improved based on more data, requiring a concerted global observational effort. The potential feedbacks on increasing atmospheric CO2 from changes in carbon flows in the coastal ocean could be disproportionately higher than in the open ocean. The changes in carbon flows and related potential feedbacks in the coastal ocean could be driven by 3 main processes: i) changes in coastal physics; ii) changes in land-used, waste water inputs, agricultural fertilizers and changes in hydrological cycle; iii) changes in seawater carbonate chemistry (ocean acidification). These potential feedbacks remain largely unquantified due to a poor understanding of the underlying mechanisms, or lack of modelling to quantify them. Based on reported evaluations and back of the envelop calculations, it is suggested that changes of biological activity due the increased nutrient delivery by rivers would provide by 2100 a negative feedback on increasing atmospheric CO2 of the order of magnitude of the present day sink for atmospheric CO2. This negative feedback on increasing atmospheric CO2 would be one order of magnitude higher than negative feedback due to the decrease of either pelagic or benthic calcification related to ocean acidification, and than the negative feedback related to dissolution of CaCO3 in sediments. The increase of export production could also provide a significant feedback to increasing atmospheric CO2, although based on the conclusions from a single perturbation experiment. Feedbacks on increasing atmospheric CO2 due to effects of C cycling in continental shelf seas related to changes in circulation or stratification could be important but remain to be quantified.
