Potential and limitations of New Zealand's pre-deforestation fossil pollen records as recent analogues in palaeoecological research

Abstract

Modern pollen analogues act as reference samples (calibration datasets), representing the observable range of vegetation types and their climatic drivers. However, in some areas, human activities have decoupled vegetation-climate relationships, resulting in modern pollen spectra that do not accurately represent those processes and are poor analogues to fossil pollen. Where these issues arise, recent fossil pollen may be an appropriate substitute. This approach is particularly relevant in New Zealand, where indigenous forest cover has been radically reduced from 85 – 90% to 25% over c. 750 years, leaving deforested areas where modern pollen samples are overwhelmed by anthropogenic signals. The point in time just prior to human arrival in New Zealand (c. CE 1280), is therefore an essential reference point for building representative calibration datasets and accurate palaeoecological reconstructions. In this review, we took a systematic quantitative approach to (1) compile a site inventory of New Zealand's pre-deforestation fossil pollen datasets from published literature (including theses), (2) identify biases and gaps in their spatial distribution (geography, climate, remaining vegetation, and potential vegetation) and describe their influence on the accuracy of future analogue-based research, and (3) comment on the availability of datasets in existing palaeoecological archives. We identified 275 datasets across New Zealand and – while they had an impressive geospatial and climatic range – uncommon vegetation types, cold and dry eastern areas, and forested reference sites were under-represented. Few raw datasets (n = 10) were readily accessible through public data repositories, and we believe New Zealand would benefit from a national palaeoecological database.