Sedimentary proxy records (e.g. speleothems, ocean sediments) are vital climate archives of the past, well distributed globally. Thus, are valuable for intra- and intercontinental assessment of past global changes. One of the key tools in exploring the climate signal is spectral analysis. However, the timescale error and variable temporal resolution of the proxy may bias the spectral characteristics.
The aim of the project is to stochastically model the spectral bias caused by timescale error on simulated time series resembling the characteristics of real-life sedimentary proxy records.
To achieve the aims, an annually sampled gap- and timescale error-free time series are used. These are resampled in a controlled way to simulate different time resolutions, and timescale error was added to it, to mimic a sedimentary proxy record with chronological uncertainty. To do so, the actual historical date of the record is the expected value and the standard deviation the arbitrary chosen uncertainty of the timescale. An ensemble of potential timescales is retrieved and their spectral characteristics were explored. These steps are taken for different resolutions as well.
It is expected that for spectral analyses the resolution is the bottleneck of the analysis, while for coherency tests both resolution and the timescale error are the ones determining the methodological boundaries.