Age‐depth models for tropical marine hemipelagic deposits improve significantly when proxy‐based information on sediment composition is included.

Accurate age-depth models for marine sediment cores are crucial for understanding of paleo-oceanographic and -climatic changes derived from these archives. To date, information on bulk sediment composition is largely ignored as a potential source of information to improve age-depth models. Here, we explore how bulk sediment composition can be used qualitatively to improve age-depth models. We developed the BomDia, algorithm, which produces age-depth models with realistic sediment accumulation rates that co-vary in harmony with the bulk sediment composition. We demonstrate that changes in the marine versus terrigenous sediment deposition based on bulk sediment composition, can be used to significantly improve age-depth models of hemipelagic marine deposits. Based on two marine records - each containing more than twenty radiocarbon (AMS 14C) dated levels - we show that the mean error of prediction of unused AMS 14C ages significantly improves from 3.9% using simple linear interpolation, to 2.4% (p = 0.003) when bulk sediment composition is included. The BomDia age modelling approach provides a powerful statistical tool to assess the validity of age control points used and also may assist in the detection of hiatuses. Testing and further development of the BomDia algorithm may be needed for application in other than hemipelagic depositional settings. Key Points The accuracy and predictive quality of age-depth models for marine hemipelagic cores improves when bulk sediment composition is included The BomDia algorithm produces age-depth models with realistic accumulation rates which co-vary with bulk sediment composition The Bomdia algorithm uses a Monte Carlo approach to assess age-depth model uncertainty Plain Language Summary The age-depth relationship in a marine sediment core is known as an age-depth model. To build an age-depth model, that is to produce a continuous age-depth relationship, an accurate and precise prediction of ages in between age control points is needed. A new algorithm, which includes information on the bulk sediment composition in the prediction of ages, has been developed and tested for this purpose. The calcium carbonate content of the sediments is calibrated versus the log ratio of titanium and calcium as obtained from resolution X-ray fluorescence and used as model input for bulk sediment composition. We tested the algorithm using data from two tropical hemipelagic cores and find that the algorithm, which we refer to as BomDia (‘a good dynamic interpolation algorithm’), produces robust age-depth models with predictive power and realistic sedimentation rates that co-vary with sediment composition. Further application and testing of the algorithm to a wider variety of other than tropical hemipelagic sedimentary settings is recommended