Bridging the gap between the terrestrial and marine paleoclimate reconstructions

Målet med dette prosjektet er å rekonstruere overflatetemperaturen i den nordlige Nordatlanteren for de siste 2000 år tilbake i tid. To avanserte metoder brukes for rekonstruksjon: (1) "BARCAST" (Bayesisk algoritme for rekonstruksjon av klimaanomalier i rom og tid, oversatt fra engelsk) og (2) analogmetoden. Datagrunnlaget er basert på proxyer for havoverflatetemperatur, disse er samlet fra kjente databaser og klargjort. Marine proxydata har dårlig tidsoppløsning og større usikkerheter i datering sammenlignet med landdata, og det er derfor viktig å teste hvor godt metodene fungerer for disse typer data. Metodikken er testet grundig for det aktuelle havområdet og datagrunnlaget i form av pseudoproxy-eksperimenter, og utgjør tema for den første publikasjonen fra prosjektarbeidet. Resultatene viser at begge metodene er i stand til å rekonstruere havtemperaturer på en tilfredstillende måte, når data antas å være perfekt datert. Kvaliteten på rekonstruksjonen varierer med simulert støy på proxydata. Ved simulerte usikkerheter i datering er det bare metode 2 som er i stand til å rekonstruere temperaturene, på tross av lavere kompleksitet sammenlignet med BARCAST.

Outcomes for Tine Nilsen as a Postdoctoral fellow holding an RCN Mobility Grant: The research has strengthened Tine Nilsen's position within the paleoclimate community, addressing an important topic highly relevant for the wider climate community. The strengthened skills within mathematics and statistics has already proven highly valuable for the future career, as she has been selected as the top applicant and accepted a permanent researcher position within this field at the Institute of Marine Research in Norway. The scientific outcomes of the research: One scientific article is published online with the main results of the project WP1. Associated with this publication, numerical implementations for the two reconstruction methodologies are also publicly available online, written in the MATLAB environment.

This project is dedicated to the assessment and extension of two state-of-the-art climate field reconstruction (CFR) techniques, applied using terrestrial and marine proxies to construct ensembles of spatially seamless surface temperature field reconstructions in both terrestrial and marine domains in the North Atlantic realm covering the past 2000 years. Focusing on the North Atlantic domain is conditioned by an overall importance of the region for influencing the climate on the global scale. The reconstruction products are used to investigate the mechanisms and forcings of past marine north Atlantic climate on the wider climate variability over Europe. The two CFR techniques under investigation are "BARCAST" (the Bayesian Algorithm for Reconstructing Climate Anomalies in Space and Time) and the proxy surrogate reconstruction method. The model assumptions of the BARCAST CFR technique represent the most important criticality of this project, since the existing methodology so far does not take different spatiotemporal covariance structures into account. The technique will be updated and tested before usage, and the alternative proxy surrogate reconstruction method will be used regardless of outcome. A successful implementation of the project is therefore secured, and the proposed reconstruction(s) are expected to be of interest for the wider climate community. Having the oceanic and terrestrial domains integrated in the framework of the project will provide means for the analysis of past changes in marine to terrestrial teleconnections in the North Atlantic realm, as well as the assessment and improvements of existing seasonal reconstructions of major climate indices, e.g. the North Atlantic Oscillation and the Atlantic Multidecadal Oscillation for the last 2000 years.