Abstract
The use of optical proxies is essential to the sustained monitoring of dissolved organic carbon (DOC) in estuaries and coastal wetlands, where dynamics occur on subhour time scales. In situ dissolved organic matter (DOM) fluorescence, or FDOM, is now routinely measured along with ancillary water-quality indicators by commercial sondes. However, its reliability as an optical proxy of DOC concentration is often limited by uncertainties caused by in situ interferences and by variability in DOM composition and water matrix (ionic strength, pH) that are typical at the land-ocean interface. Although corrections for in situ interferences already exist, validated strategies to account for changes in the DOM composition and water matrix in these systems are still lacking. The transferability of methods across systems is also poorly known. Here, we used a comprehensive data set of laboratory-based excitation-emission matrix fluorescence and DOC concentration matched to in situ sonde measurements to develop and compare approaches that leverage ancillary water-quality indicators to improve estimates of DOC concentration from FDOM. Our analyses demonstrated the validity of in situ interference correction schemes, the importance of ancillary water-quality indicators to account for DOM composition and water matrix change, and the good transferability of the proposed methods.