Berkeley Institute for Data Science

–Interactions between biodiversity and fire in California ecosystems
Published

August 29, 2023

Project Overview

This project is part of a fellowship in Global Change Research at the Berkeley Institute of Data Science (BIDS).

As we enter an era of unprecedented amounts of data, we also face the defining challenge of our age - global environmental change. From automated environmental sensors to satellite imaging to emerging DNA technologies, the data that describes the world around us will aid us in how we approach the many challenges of a changing climate. By harnessing this data we can develop data science tools that allow us to predict, and further define how we as a society can respond and mitigate the effects of climate change.

My main project is developing a multi-level modeling framework that incorporates environmental variables, satellite data, historical species observations, and environmental DNA (eDNA) data across fire regimes in Californian ecosystems. The collated data sets and resulting modeling framework will enable researchers to better understand the impact of historical and recent fire on arthropod communities in sites across the University of California Natural Reserve System. We hope that this framework will be able to be applied to other open questions regarding fire and biodiversity. This is a unique project that is rooted in basic research but provides a framework for interoperability between different data types. The project outcomes will be disseminated through a research blog and tutorials, science communications through illustrations and zines, open-source code, and open publications.

Outcomes

  • Paper Accepted to Global Change Biology Holmquist AJ, Markelz RJC, Martinez CC, Gillespie RG. The importance of habitat type and historical fire regimes in arthropod community response following large-scale wildfires. doi: https://doi.org/10.1101/2023.07.17.548903 . Available for comment here.

  • Open-source data pipeline for analysis: The code repository is available here. The repository also contains instructions on how to make reproducible data science docker containers for each of the major steps in the analysis pipeline. We made the docker container with the largest number of software dependencies available on docker hub here.

  • Data Science by Design (DSxD) Anthology Volume 2: Our Environment. Fire Zine published in printed anthology. Learn more here

  • Book Chapter Co-author and Back Cover Illustration - Van Tuyl, Steve (Ed.). (2023). Hiring, Managing, and Retaining Data Scientists and Research Software Engineers in Academia: A Career Guidebook from ADSA and US-RSE. Zenodo. https://doi.org/10.5281/zenodo.8274378

  • Teaching: Field fire ecology at Merrit Community College using zines and illustrations. I recently blogged about a introductory fire ecology class at the Blue Oak Ranch Reserve. The reserve is part of the UC Reserve system and is one of the many sites I am collecting data at for a larger project looking at fire impacts on California ecosystems and insect pollinators. Other blog posts in the series are also available: post 2, post 3, post 4. I wrote about the reserve system in these other posts on fire data and environmental data.

  • Collated data resources page: Fire Data Resources.

Ongoing

Pyrodiversity of the Klamath Mountains

  • Extend research findings through blog posts about fire in California’s Klamath Mountain Range that will be recombined to make up a feature length general science article. Other products: data visualization, data storytelling, science zines
  • Wildland Urban Interface in California - fire science and climate impacts. Products: data visualization, data storytelling, science zines
  • Post-fire impacts on aquatic insect biodiversity - Products: data visualization, data storytelling, science zines

Everyday Data Visualization Initiative

  • Create an open-source book (using www.quarto.org), code, and graduate/postdoc training workshops centered around collecting, summarizing, modeling, and visualizing data that is collected from common devices and applications.
  • Personal data source examples include fitbit/apple watch movement data, geotagged meta-data of images on phones, and social media network data.
  • Publicly available data would include weather, climate, traffic, or map data.
  • Each lesson will be written as a series of blog posts for content creation and shared through social media channels as they are developed.

Mentors

  • Ciera Martinez, Biodiversity and Environmental Sciences Lead, BIDS
  • Rosemary Gillespie, Professor, Environmental Science, Policy, and Management
  • David Ackerly, Dean, College of Natural Resource
  • Karthik Ram, Senior Research Data Scientist, BIDS
  • Accenture Applied Intelligence, to fully harness the data landscape and expertise available. Through independent fellowships, Accenture supports BID’s research and educational objectives in data science with current foci on environment and energy, ethical AI, and social justice.