Sensitivity of sediment geochemical proxies to coring location and corer type in a large lake: Implications for paleolimnological reconstruction

We compared a suite of geochemical proxies in sediment cores collected in 1982, 1988, 1991, and 2003 from sites near the depocenter of Lake Erie to evaluate the reliability of paleoenvironmental reconstructions derived from lacustrine sediments. Our proxies included the concentrations and carbon isotopic compositions of organic and inorganic carbon (TOC, CaCO3, d Corg, and d CCaCO3), augmented by organic C to total N ratios (Corg:Ntot), d N, and carbonate dO values (dOCaCO3). The three coring sites were clustered within 12 km; two types of corers—a Box corer and a Benthos gravity corer—were used for the 1991 sampling campaign. The variance of most proxies was accounted for not only by temporal environmental changes but also by coring locations and corer type, indicating that sediment spatial heterogeneity and differences in sediment recovery due to the use of different corers also played a part in determining the geochemical compositions of these cores. The TOC, dCorg, and d CCaCO3 values showed decadal temporal patterns that were consistent between the multiple sampling campaigns. In contrast, the dN, Corg:Ntot, CaCO3, and d OCaCO3 exhibited across-core differences in their temporal variations, making it difficult to extract consistent environment information from different cores. Our findings suggest that in addition to temporal environmental changes, high-resolution paleolimnological reconstruction is sensitive to many factors that could include spatial sediment heterogeneity, discontinuous sedimentation processes, bioturbation, sediment dating uncertainty, and artifacts associated with analytical and coring procedures. Therefore, multiple-core sampling and analysis are important in reliably reconstructing environmental changes, particularly for large, heterogeneous lacustrine basins.