LinkedEarth at the AGU Fall Meeting 2021

Deborah Khider
CyberPaleo
Published in
5 min readJan 24, 2022

TL;DR: Our first community town hall took place at the American Geophysical Union Fall Meeting 2021 in New Orleans, where we announced a series of new events for the year, listened to a presentation by Dr. Georgy Falster (Australian National University) and answered community questions. Slides from the presentation are available here.

Read on for a summary of these annoucements.

LinkedEarth at AGU

We would first like to thank all of you who attended our first (which should have been hybrid) town hall at AGU in New Orleans. We must apologize for the technical difficulties.

During the hour-long town hall, we announced new initiatives (detailed below), demoed tools that were developed as part of several LinkedEarth activities (namely, the LiPD playground, the LiPDverse, GeoChronR, and Pyleoclim), listened to talk by Dr. Georgy Faster (Australian National University) analyzing the ISO2k database using LinkedEarth tools, and answered questions from the community.

The topics ranged from data stewardship, analysis, data products, and community initiatives.

Data Stewardship

Most of LinkedEarth’s efforts so far has been in data stewardship with the introduction of the Linked Paleo Data (LiPD format) and a reporting standard (PaCTS). In order to facilitate adoption of this standard by the community, Nick McKay worked relentlessly to create a GUI for the annotation of LiPD files (the LiPD playground) and validation against several repositories. The playground has recently been augmented with a recommender system to assist with metadata annotation. In the next year, we are planning to add a feature that would allow a user to send a dataset from their Python workspace to the playground for viewing/editing and send it back to Python for analysis. We are also planning to align the vocabulary in the LinkedEarth Ontology to the PaST Thesaurus developed by NOAA.

Screenshot of the LiPD playground.

Another tool is the LiPDverse, which allows access to thousands of curated LiPD files and many compilations. In the next few months, look for some website redesign and tutorial blogs. In the coming year, we will also enable querying directly from Python and R.

Analysis

Most of our work will focus on creating analysis tools that make use of these standards to make the work of paleoclimatologists easier. The first step is to make the LiPD utilities friendlier in Python. In addition, we will automate the import of data stored in NOAA WDS Paleo and PANGAEA into a LiPD structure. The structure can then be sent to the playground for editing if needed.

In the spring of 2021, the GeoChronR manuscript was published, detailing the capabilities of the package as well as examples of scientific applications. Throughout 2022, we will also release a series of short tutorials, tools for multi-model integration, and other user requests. If you have a request for the package, submit them as GitHub issues!

GeoChronR vignette

The biggest improvement will be to the Pyleoclim software package, with better integration with Pandas and xarray. If you are interested in learning more about the development of Pyleoclim, feel free to join our development meetings held virtually every other Tuesdays from 1–2pm Pacific (subscribe to our calendar for the meeting link). Requests for features can also be made on our GitHub repository.

Pyleoclim dashboard

You will soon be able to work in a dedicated cloud environment with the package. In partnership with 2i2c, we will be deploying a JupyterHub with available compute nodes. This is your space! You can create your own code and upload your own data; they will not be shared with anyone else. All you need is a GitHub account.

To help you get started with your own notebooks, we will also soon launch a gallery of Jupyter Notebooks, working through scientific examples. Our focus will be on paleoclimate model/data comparison in the time and frequency domain. These notebooks will be fully executable through myBinder and on our JupyterHub. If you are interested in contributing to the gallery, please let us know! Although we would love for you to use our tools, the gallery is open to any projects coded in Python.

During the hackathon, Nick and Julien demoed GeoChronR and Pyleoclim. The video tutorial for Pyleoclim is available on our YouTube channel.

Data Products

One of our newer projects is the Paleoclimate Reconstruction Storehouse (PReSto), a one-stop shop for paleoclimate field and index reconstructions. Reconstructions will be regularly and automatically updated. Users will be able to visualize data directly on the platform and download existing reconstructions as well as making their own reconstructions using data from published compilations (e.g., PAGES2k, Temp-12k) and published algorithm or bring their own data/methods.

Community

Chiefly, LinkedEarth is about creating a community of paleogeoscientists sharing their data, methods, scientific findings, and other insights. To this end, we are enabling more and more community activities.

We have created a Discourse forum for discussion about anything paleo: any news and announcements you might have, discussion about data and software (and the standards that support them), training opportunities, science, visualization, education, job opportunities, and JEDI.

We will also be releasing tutorials on LiPD, GeoChronR, and Pyleoclim through our YouTube channel and as blogs on Medium.

In addition to tutorials, we will also be organizing training (virtual) hackathons for hand-on experience with LiPD, GeoChronR, and Pyleoclim. We envision three types of hackathons: beginner (no prior experience with the tool or R/Python required; go through introductory materials/basic methodologies for time series), intermediate (for the R/Python savvy and people with experience with Pyleoclim/GeoChronR; work through science problems), advanced (bring your own data/question or help develop a new feature in GeoChronR/Pyleoclim).

Our next hackathons will be held in the Spring 2022:

  • GeoChronR: February 8–9th, apply here.
  • Pyleoclim (Beginner): March 9th-11th, apply here.

If you want to get more engaged with LinkedEarth, please attend our community meetings (just tell us about the great science you’re doing!), every other Monday 9am to 10am. Details in our calendar.

If you have any feedback about the town hall and if you’d like to see more events like this in the future, please let us know!

Deborah, on behalf of the LinkedEarth team.

--

--

Deborah Khider
CyberPaleo

Research Scientist at the USC Information Sciences Institute - Data Science, AI, and paleoclimatology