Modeling catchment controls on organic carbon fluxes in a meso-scale mountainous river (Luodingjiang), China

This study applied a partial least square regression (PLSR) technique to investigate the relationships between organic carbon fluxes and catchment environmental factors in the Luodingjiang, a meso-scale mountainous tributary of the Zhujiang (Pearl River), China. Catchment environmental characteristics were extracted from various ecological datasets, and river data were developed based on 13 sampling points during six sampling campaigns in 2005. Modeling results indicate that the lithology characterized by rock erodibility in the catchment is the most important control on riverine organic carbon fluxes (dissolved organic carbon (DOC) flux, and particularly particulate organic carbon (POC) flux) and total suspended sediment (TSS) flux in the Luodingjiang. Besides rock erodibility, more environmental variables are positively involved in the model of DOC flux, including precipitation, urban and agricultural area percentages, as well as basin area, while basin slope is negatively involved. Soil organic carbon (SOC) content does not exert a significant control on DOC flux in the study due to its unique spatial distribution with other environmental factors having reverse relationships with DOC flux, such as basin morphology and anthropogenic impacts. The dominant control of lithology on POC flux and TSS flux is considered as a result of the availability of highly erodible rocks in the region, such as coarse-grained granite and poor-cemented purple sandstone. The models developed in this study may be only applicable in the local catchments with similar environmental characteristics, such as hilly regions with weak underlying bedrock. The findings highlight the need for investigating the geographical variations when conducting studies about riverine organic carbon fluxes at the regional or global scale.