COS 63-5 - Reproducible Reproducibility: Open Research Workflows for Ecosystem Science

Abstract: The toolkit for open and reproducible [ecosystem] science is expanding rapidly (e.g., computational reproducibility, clean codes). There is a strong emphasis on reproducibility as the ability to iterate workflows and pipelines to re-produce a given research object (e.g., data analysis or an entire research paper). This product-driven approach appears highly tailored for most scientists, at the risk of excluding other professional disciplines and creating entry-level barriers to collaboration and engagement. Here, we build on the knowledge management systems theory to generalize reproducibility workflows and their implementation via computer code (i.e., pipelines) and leverage these technical skills for science communication and engagement. Our workflow consists of five (5) interconnected building blocks that describe a modern system for knowledge production: 1) Knowledge bases, 2) Analytical engines, 3) Production Hubs, 4) Product stores, and 5) User interfaces. The workflow is implemented with the latest pipelines for computational reproducibility and user engagement (RMarkdown, Quarto, and GitHub Pages). We present examples of specific implementations of this workflow to create reproducible data downloads, ecological literature reviews, shared text- and data analysis, and collaborative ecological syntheses. Our resulting workflow is highly customizable and device-agnostic and allows for version control across the entire knowledge production system: from conversations to publications to science communication. As the ecological challenges we face increase in complexity, we want to ensure that the science we will pass on to the next generation will carry the hallmarks of openness and reproducibility and the blueprint to create open and reproducible science.