Particulate organic carbon (POC) sources, which regulate dissolved organic carbon, sediment organic carbon, and inorganic carbon via deposition, degradation, and mineralization, play an important role in lake ecosystems. Linear or Bayesian algorithms on isotope and n-alkanes have been widely used to identify the source proportion of organic carbon. However, the applicability of these methods is ambiguous because of the unilateral advantages of each model and trace factors. To test the applicability of the various methods for identifying POC sources, we analyzed dual isotopes and n-alkanes in surface water samples of Lake Taihu, and Multi-source mixing model and Bayesian mixing model were used to distinguish between endogenous and exogenous contributions. Carbon isotope presented a clear advantage in West Taihu (-21.85 ± 0.78‰) and Southwest Taih (-22.61 ± 1.35‰); nitrogen isotope also showed high values in Meiliang Bay (9.76 ± 0.92‰). The majority of the lake was dominated by short-chain n-alkanes, except for East Taihu Lake (dominated by medium-chain n-alkanes) and areas with riverine input (dominated by long-chain n-alkanes). Different principles between the Bayesian mixing model (based on the Markov Chain Monte Carlo algorithm) and the Multi-source mixing model (based on linear estimation) caused discrepancies in the estimations of source contributions. But the fraction of chemical compounds during the migration process, and the overlap of potential sources play important role in the inconsistency of results. The estimations from the different models were consistent in indicating the dominance of endogenous organic carbon in Lake Taihu (mean of 60.18 ± 20.26%), particularly in the north and western regions (West Taihu, Meiliang Bay, and Southwest Taihu). This was likely due to algal aggregation influenced by human activities and climatic factors.
Keywords: Algal blooms; Bayesian mixing model; N-alkanes; Particulate organic carbon; Stable isotope.
Copyright © 2021. Published by Elsevier Ltd.