It has been established that coral reef flats are mainly sources of atmospheric CO 2 due to the high biocalcification of coral reef communities in relation to low net fixation of CO 2 via photosynthesis (Gattuso, Frankignoulle, et al., 1999 and references therein). Recent observations in Great Barrier Reef and Hawaii tropical coral reefs confirm this overall trend (Lønborg et al., 2019;Terlouw et al., 2019), although localized processes make the drivers more complex. The global CO 2 emission from coral reefs is roughly estimated at 7 Megatons of carbon per year, based on field measurements from a "reference" reef in Moorea, French Polynesia (Gattuso et al., 1993;Spalding et al., 2001). It is worth noting that approximately 0.6 mol CO 2 was released when 1 mol CaCO 3 precipitates over a "standard" reef flat, where net community production (NCP) was close to zero in a daily cycle (Kinsey, 1983;Ware et al., 1992). However, it is possible that the release of CO 2 could be much lower, or even negative, in algae-dominated coral reefs, which may exhibit a "sink" pattern for atmospheric CO 2 (Frankignoulle Abstract CO 2 signatures differ among locations, even within the same coastal habitat. However, the driving mechanism behind this variation is not well understood due to the presence of both human and natural stresses. This study examines seawater pCO 2 and its deviation from the regional average (defined as pCO 2 anomaly) during spring, summer, and winter at two contrasting zones (trial and core zones) of Dongshan coral habitat. The impact of solar, tidal, and biological cycles on pCO 2 anomaly is relatively small under the designed sampling scheme. Results show that there is a large positive pCO 2 anomaly in the turbid water of the trial zone during all studied seasons, while the large negative pCO 2 anomaly in the core zone only occurs during the summer upwelling season. This indicates that CO 2 signatures among locations may have seasonality. The physical and biological contributors to pCO 2 anomaly are then quantitatively revealed based on a first-order Taylor decomposition and biological carbon metabolisms. The results suggest that organic carbon metabolism of marine phytoplankton dominated pCO 2 anomaly in all three seasons, while inorganic carbon metabolism also enlarged pCO 2 anomaly, especially during summer. Physical contributions were minor. Moreover, significant linear correlations between turbidity (related to local human activities), chlorophyll (related to coastal upwelling) and changes in dissolved inorganic carbon (induced by organic carbon metabolism) indicate human and natural influences on bioprocesses, which may enhance pCO 2 anomaly. Finally, the Dongshan coral habitat was estimated as a source of atmospheric CO 2 , with an annual efflux of approximately 50 t C.
Plain Language SummaryPrevious studies have shown that CO 2 levels in seawater vary depending on location, and even within the same habitat. However, it is still unclear why this happens, as human activities and natural processes overlap...