Use of synchronous fluorescence spectra to estimate biochemical oxygen demand (BOD) of urban rivers affected by treated sewage

Synchronous fluorescence spectra and the first derivative spectra of urban rivers affected by treated sewage were studied to find the optimum index for biochemical oxygen demand (BOD) estimation. In addition, various sample treatment methods as well as the multiple regression method using available monitoring parameters were examined to improve the BOD estimation capability of the indices. From July to November 2006, grab samples were collected from three rivers in the city of Seoul, Korea, which included one receiving effluent from a municipal wastewater treatment facility. The BOD values of the samples ranged from 0.4 to 28.0 mg l(-1). Out of the many fluorescence characteristics observed, the relative fluorescence intensity at 283 nm to 245 nm from the synchronous spectra was selected as the optimum fluorescence estimation index because it had the highest positive correlations with the BOD values (r = 0.910). The filtration process, using GF/F filter prior to the fluorescence measurement, resulted in an enhanced correlation demonstrating that the removal of suspended solids from the samples helps alleviate, to some extent, the interference of light scattering. In contrast, the same correlation was not enhanced by the pH adjustment of the sample to minimize the complexation of organic matter with cations. The multiple regression method, using either light scattering intensity at 633nm or turbidity, greatly enhanced the correlation between measured and predicted BOD values. The correlation was even higher than that obtained for filtered samples, thus suggesting that the multiple regression method based on the original sample is more effective than sample treatment in enhancing BOD estimation.