Abstract
The ability of absorbance at 330 nm to predict dissolved organic carbon (DOC) concentration was tested on samples of throughfall, stem‐flow, soil water, and stream water from the Maimai and Larry River catchments, Westland, New Zealand. Predictive ability for the combined samples was poor, but improved when the samples were split into groups of origin. For stream samples, the method gave standard errors of the estimate of 8 to 11% of the mean. Predictive ability was poorer in the other sample groups, especially throughfall and stemflow, presumably because of the occurrence of weakly coloured organic compounds. The slope and intercept of the regressions varied with sample type. The spectrophotometric method is useful where a large number of samples need to be analysed and where calibrations can be developed for specified types of organic‐carbon‐rich waters.