Multi-tracer evidence for the presence of autochthonous organic carbon and the role of biological carbon pump in two river–reservoir ecosystems on the Chinese Loess Plateau

The biological carbon pump (BCP) effect of dissolved inorganic carbon (DIC)-enriched karst surface waters is an important carbon sink mechanism that has substantial implications for the study of global missing carbon sinks and carbon cycling. However, the operation of the BCP is not established well in wider inland waters, particularly not in the surface waters of the Chinese Loess Plateau (CLP) that are also DIC-enriched but have high turbidity. In this study, we combined analyses of stable isotopes (δ²H, δ¹⁸O, and δ¹³C) and C/N ratios, ultraviolet absorption spectroscopy, and excitation-emission matrix fluorescence approaches to investigate the sources and compositions of particulate organic carbon (POC) and dissolved organic carbon (DOC) in two river–reservoir ecosystems [the Wulihe (WLH)-river to WLH-reservoir and Honghe (HH)-river to HH-reservoir] on the CLP. Both DOC and POC signals from autochthonous production gradually increased from the river inflow to the reservoir areas along the river flow, particularly during the wet season with the same duration of rain and heat. The WLH-reservoir had higher δ²H, δ¹⁸O, and δ¹³C values, more protein-like fluorescence components, higher chlorophyll-a and DOC concentrations, and lower CO₂ efflux than the HH-reservoir. Our results showed that the BCP intensity of the WLH-reservoir was higher than that of the HH-reservoir, probably because of the longer water residence time of the WLH-reservoir. Although the amount of autochthonous dissolved organic matter (Auto-DOM) produced by the BCP in the surface water on the CLP was lower than that in the karst regions of southwest China (CKR), the proportion of Auto-DOM in the CLP was close to that in the CKR, indicating the significance of Auto-DOM in the surface water on the CLP. Furthermore, the BCP-derived Auto-DOM correlated negatively with the partial pressure of CO₂, and the WLH-reservoir with a higher BCP effect exhibited CO₂ influx in the wet season (−1.47 ± 0.02 g m⁻² d⁻¹), indicating that the enhanced BCP process contributed to the water–air interface carbon sink. This study highlighted that enhanced BCP could generate carbon sinks in DIC-enriched but high-turbidity surface waters, providing the underlying theory for extending the BCP carbon sink model to a wider range of surface waters flowing through areas of soil carbonate.