Integrated approach for quantitative estimation of particulate organic carbon sources in a complex river system

Despite receiving a considerable amount of attention in the past, quantitative and systematic estimation of the source contributions for different organic carbons (OCs) in complex river systems is still challenging. In this study, we tested an integrated framework using field data of bulk elements and lipid biomarkers and hydrological modeling (hydrological simulation program FORTRAN, HSPF) for the quantitative estimation of OC loads along different land-use types of a watershed (Geumho River watershed in South Korea). Based on the specific source assignments identified from the lipid biomarker patterns in particulate organic carbon (POC) such as short/long chains of alkanes, fatty acids and alcohols, and coprostanol/cholesterol, spatial variations of the diagnostic lipids could be used as an indicator to discriminate between the contributions of natural (algae, bacteria, and terrestrial plants) and anthropogenic sources (fecal). Based on the integration of HSPF modeling, it was also found that various POC loads might be partially controlled by different water discharges within watersheds. With the increase in POC fluxes, the increase in fecal loads was also noticed, as reflected by the predominant lipid (especially coprostanol normalized by water discharges). As a straightforward approach, we developed a set of indices including fecal index-1, ratios of coprostanol, fatty acids, and alkanes, which strengthened the sensitivity for fecal contamination. Compared with the conventional HSPF results, the variations of these proposed indices were more influenced by the broadened watershed extents with increasing downstream distance, which provided a more accurate estimation of the quantitative contributions of POC loadings in the complex river system.