[1] The influence of mangrove forests on the dynamics of dissolved inorganic carbon (DIC) in tropical estuaries was estimated quantitatively using newly developed isotope (d 13 C) mass balance models that take into account both the input of DIC and the air-water exchange of CO 2 . To this aim, the concentration and d 13C of DIC were determined across the salinity gradient of two river estuaries facing the Andaman Sea. The longitudinal distribution of DIC could be explained by conservative mixing of the river water and seawater DIC in the low-discharge period (March 2006), while a net accumulation of DIC up to 190 mmol L À1 was observed in the high-discharge period (December 2006). d 13 C DIC values were generally lower than expected for the mixing of the river water and seawater DIC, due to the 13 C-depleted DIC inputs from the riverside mangroves. The concentration of mangrove-derived DIC in the estuarine waters was estimated by the proposed models to be as much as 856 mmol L À1 , and was higher during the low-discharge period. This suggested that the mangroves exported much higher levels of DIC to the estuaries than indicated by the net accumulation of DIC. Our results confirm that mangroves function as an effective CO 2 pump that takes CO 2 from the atmosphere and releases it into estuarine waters. This study illustrates that d 13 C DIC is a sensitive and quantitative indicator for DIC emission to the sea from coastal wetlands including mangroves.

The longitudinal variations in the nitrogen (d 15 N) and oxygen (d 18 O) isotopic compositions of nitrate (NO 3 -), the carbon isotopic composition (d 13 C) of dissolved inorganic carbon (DIC) and the d 13 C and d 15 N of particulate organic matter were determined in two Southeast Asian rivers contrasting in the watershed geology and land use to understand internal nitrogen cycling processes. The d 15 N NO 3 became higher longitudinally in the freshwater reach of both rivers. The d 18 O NO 3 also increased longitudinally in the river with a relatively steeper longitudinal gradient and a less cultivated watershed, while the d 18 O NO 3 gradually decreased in the other river. A simple model for the d 15 N NO 3 and the d 18 O NO 3 that accounts for simultaneous input and removal of NO 3 -suggested that the dynamics of NO 3 -in the former river were controlled by the internal production by nitrification and the removal by denitrification, whereas that in the latter river was significantly affected by the anthropogenic NO 3 -loading in addition to the denitrification and/or assimilation. In the freshwater-brackish transition zone, heterotrophic activities in the river water were apparently elevated as indicated by minimal dissolved oxygen, minimal d 13 C DIC and maximal pCO 2 . The d 15 N of suspended particulate nitrogen (PN) varied in parallel to the d 15 N NO 3 there, suggesting that the biochemical recycling processes (remineralization of PN coupled to nitrification, and assimilation of NO 3 --N back to PN) played dominant roles in the instream nitrogen transformation. In the brackish zone of both rivers, the d 15 N NO 3 displayed a declining trend while the d 18 O NO 3 increased sharply. The redox cycling of NO 3 -/NO 2 -and/or deposition of atmospheric nitrogen oxides may have been the major controlling factor for the estuarine d 15 N NO 3 and d 18 O NO 3 , however, the exact mechanism behind the observed trends is currently unresolved.