LinkedEarth Lectures

The future of old things: geoinformatics for better paleoscience

Talk given by Julien Emile-Geay at the PAGES Open Science Meeting in Zaragoza, Spain,9-13 May 2017

Abstract: By some accounts, paleoscientists spend up to 80% of their time trying to access the data they need, in the form they need it. In the 21st century, we should be able to do much better. This lecture will review recent progress made by the LinkedEarth project, which relies on data standards and artificial intelligence to enable scientists to spend more time doing the science they want to do. LinkedEarth is manifesting a better future for paleoscience by creating an online platform that (1) enables the curation of a publicly-accessible database by paleoclimate experts themselves, and (2) fosters the development of community standards. In turn, these developments enable cutting-edge data-analytic tools to be built and applied to a wider array of datasets than ever possible before, supporting more rigorous assessments of the magnitude and rates of pre-industrial climate change. We will start by illustrating these principles in the context of the PAGES 2k project, and outline how they may serve the PAGES community as a whole. In particular, we will illustrate how to go from spreadsheets to syntheses (PAGES 2k). We will dwell on community participation in the first paleoclimate data standard. We will present GeoChronR and Pyleoclim, new open-source tools compatible with these standards and enabling cutting-edge paleoscience. We will finish by some remarks on interoperability, enabling cross-talk between scientists within a field, across fields, and between data and models. In our vision of the future, machines serve scientists, not the other way around. Yet, the process needs a lot of human input, and the participation of the PAGES community will be recognized and further encouraged.

You can access the slides here

 

The Future of Past Climates: EarthCube and 21st century paleoclimatology

Paleoclimate observations are crucial to assessing current climate change in the context of past variations. However, these observations often come in non-standard formats, forcing paleogeoscientists to spend a significant fraction of their time (by some accounts, up to 40%) searching for and accessing data, or converting between formats before they can do science. This considerable waste of resources hinders re-use and hence lowers the value of the datasets to scientists and society alike. In the 21st century, we should be able to do much better.

The EarthCube-supported LinkedEarth project aims to manifest a better future by creating an online platform that (1) enables the curation of a publicly-accessible database by paleoclimate experts themselves, and (2) fosters the development of community standards. In turn, these developments enable cutting-edge data-analytic tools to be built and applied to a wider array of datasets than ever possible before, supporting more rigorous assessments of the magnitude and rates of pre-industrial climate change. 

In this talk, I will first provide an example of how paleoclimate observations can be used to characterize the uniqueness of 20th century climate change in the context of the past two millennia, using novel statistical approaches. Using a sea surface temperature (SST) composite record from eight high-resolution sedimentary records from the Indo-Pacific Warm Pool, I will show that, in this region, there is a >60% probability that the late 20th century is unprecedented in the last 2,000 years. I will then show how the LinkedEarth platform can help put those tools in the hands of every paleoclimatologist, enable more efficient data curation via artificial intelligence, and support broader data syntheses that easily integrate all published information. In this vision of the future, machines serve scientists, not the other way around. 

Watch it here