Simple noise estimates and pseudoproxies for the last 21k years

Abstract

Abstract. Climate reconstructions are means to extract the signal from uncertain paleo-observations, i.e. proxies. It is essential to evaluate these to understand and quantify their uncertainties. Similarly, comparing climate simulations and proxies requires approaches to bridge the, e.g., temporal and spatial differences between both and address their specific uncertainties. One way to achieve these two goals are so called pseudoproxies. These are surrogate proxy records within, e.g., the virtual reality of a climate simulation. They in turn depend on an understanding of the uncertainties of the real proxies, i.e. the noise-characteristics disturbing the original environmental signal. Common pseudoproxy approaches so far concentrated on data with high temporal resolution from, e.g., tree-rings or ice-cores over the last approximately 2,000 years. Here we provide a simple but flexible noise model for potentially low-resolution sedimentary climate proxies for temperature on millennial time-scales, the code for calculating a set of pseudoproxies from a simulation and, for one simulation, the pseudoproxies themselves. The noise model considers the influence of other environmental variables, a dependence on the climate state, a bias due to changing seasonality, modifications of the archive (e.g., bioturbation), potential sampling variability, and a measurement error. Model, code, and data should allow to develop new ways of comparing simulation data with proxies on long time-scales. Code and data are available at https://doi.org/10.17605/OSF.IO/ZBEHX.