Join us Monday 8/22 at ESIP Cloud Computing Cluster Session on Metadata (STAC) modeling of Forecasting Data

This month’s ESIP cloud computing virtual session will feature Carl Boettiger, who comes to discuss the Ecological Forecasting Initiative’s cloud computing infrastructure.

Please see the abstract below, start pondering on your experience in the area, and come
ready to talk!

Meeting Logistics!
Topic: Ecological Forecasting Initiative on-going Forecasting Challenge,
Goals, and Infrastructure
Invited Speaker: Carl Boettiger
When: Monday August 22nd, 10:00-11:00 am PT / 1:00-2:00 pm ET

Meeting ID: 865 3517 7705
Passcode: 354962

The Ecological Forecasting Initiative (EFI) challenge seeks to
facilitate ecologists producing public, iterative ecological forecasts by
using a simple, standards-based cloud-native infrastructure. The current
challenge focuses on five themes in data products produced by the National
Ecological Observatory Network (NEON), including community ecology,
population dynamics, CO2 flux, aquatic ecosystem indicators, and phenology.
Ecological models are nearly always built in an artisanal, hand-crafted
fashion, which means that designing forecasts which can be automated and
scaled up to available data is a significant hurdle for most of our
intended audience. However, by leveraging bucket-based object stores (such
as AWS S3 or open source equivalents) and recent developments supporting
cloud-native storage in cross-language tools such as Apache Arrow, we have
been able to provide a simple, scalable approach for teams to easily access
NEON data, NOAA forecasts, submit their forecasts to the challenge and
visualize their accuracy compared to competing teams as new data becomes
available. In the process, we’ve learned some important lessons about
building a low cost, low maintenance, accessible cloud-based infrastructure
for research teams.


  • 5-10 minutes - Cluster and ESIP Announcements, open call for other
  • 10-15 minutes - Problem statement and lay of the landscape
  • 15-35 minutes - Open discussion
1 Like
1 Like