A guide to assess distance from ecological baselines and change over time in palaeoecological records

Palaeoecology provides critical data for establishing ecological ‘baselines’, which can guide restoration efforts and be used to assess ecosystem change. However, statistical analyses can be challenging because of the large number of methods available for establishing palaeoecological baselines combined with a lack of practical guidance, particularly around quantifying baselines that include natural variability. We contribute one solution by providing guidance and an R package baselines for using palaeoecological data to (i) define baselines and (ii) test for change over time that incorporates variability. These methods provide an alternative to single-taxon analyses and allow ecosystem complexity to be captured. We use published pollen records as case studies to demonstrate how to establish vegetation baselines for seven localities in New Zealand where relatively recent (c. 1280 AD) and near-simultaneous human settlement across the country allows background environmental disturbance in the pre-human era to be distinguished from anthropogenic disturbance. We present methods for calculating distance from initial sample, distance from baseline in ordination space, allowing incorporation of ecosystem variability, and analysis of rates of change over time using principal curves. We found conventional and Bayesian ordination methods yielded similar results and were effective at identifying change following human settlement, despite the potential for a positive mean-variance relationship to confound results. Principal response curves were most sensitive to a known period of vegetation disturbance caused by volcanic eruptions at two sites with tephra deposits. Our proposed methods, case study and R package baselines are designed to provide a suite of tools to encourage and enable palaeoecological data to be used by palaeoecologists to assess trajectories and change over time, and monitor whether historical management actions have facilitated a change in direction towards a desired baseline state.