Quantification of sedimentary organic carbon sources in a land–river–lake continuum combined with multi-fingerprint and un-mixing models

Abstract

Identifying organic carbon (OC) sources in lake sediment is essential for elucidating biogeochemical cycling processes and effectively supporting watershed management. However, the complexity of sources as well as environments in the land–river–lake continuum makes it challenging to accurately identify OC sources. Accordingly, the current study utilized a systematic approach to identify and validate OC sources in a typical land–river–lake continuum. Two tracer groups (group 1: δ¹³C and δ¹⁵N; group 2: fluorescence index and biotic index, respectively (where C is carbon and N is nitrogen)) and one model (MixSIAR) were eventually selected from five tracer groups and two models to identify the OC sources in a land–river–lake continuum according to a consistency evaluation and virtual mixing test. The results showed that the distribution of OC sources in lake sediment was spatially heterogeneous. Closer to the lake center (from sampling site S1 to S3), the autochthonous contributions increased while the allochthonous contributions decreased. Downstream of the inlet river (site S1) was dominated by allochthonous contributions (78.6%), especially cropland (28.7% ± 0.5%, where ± indicates a standard deviation range) and urban land (30.5% ± 2.5%). From site S1 to S2, the allochthonous contribution decreased 11.4%. Autochthonous OC gradually became the major source closer to the lake center (site S3: phragmites: 48% ± 4.5%). This distribution of OC sources in the land–river–lake system was attributed to the mixing effect of the autochthonous sources, selective transport of sediment, and human activities. The current findings may aid in validating the ability of different tracers and models to identify OC sources in complex ecosystems and also provide a theoretical basis for watershed management.