Open science, reproducibility, and transparency in ecology

TitleOpen science, reproducibility, and transparency in ecology
Publication TypeJournal Article
Year of Publication2019
AuthorsPowers, SM, Hampton, SE
JournalEcological Applications
Type of ArticleJournal Article

Reproducibility is a key tenet of the scientific process that dictates the reliability and generality of results and methods. The complexities of ecological observations and data present novel challenges in satisfying needs for reproducibility and also transparency. Ecological systems are dynamic and heterogeneous, interacting with numerous factors that sculpt natural history and that investigators cannot completely control. Observations may be highly dependent on spatial and temporal context, making them very difficult to reproduce, but computational reproducibility can still be achieved. Computational reproducibility often refers to the ability to produce equivalent analytical outcomes from the same data set using the same code and software as the original study. When coded workflows are shared, authors and editors provide transparency for readers and allow other researchers to build directly and efficiently on primary work. These qualities may be especially important in ecological applications that have important or controversial implications for science, management, and policy. Expectations for computational reproducibility and transparency are shifting rapidly in the sciences. In this work, we highlight many of the unique challenges for ecology along with practical guidelines for reproducibility and transparency, as ecologists continue to participate in the stewardship of critical environmental information and ensure that research methods demonstrate integrity.

Ecologists have long faced a novel challenge not routinely encountered in less field-oriented sciences: repeated testing is fundamental to the scientific method (Popper 1934) yet it is impossible to perfectly repeat observational studies of the natural world (Vanderbilt and Blankman 2017). This issue is timely as scientists across disciplines increasingly recognize the challenges of reproducing published results, and the threats that irreproducible results pose to the scientific process (Munafo et al. 2017). Reproducibility and transparency issues are particularly important for scientists engaged in actionable science and ecological applications; this work often feeds back rapidly and directly on biota, ecosystems, and people who have stakes in conservation or management outcomes, in turn affecting perceptions about the integrity of our field.

For ecologists engaged in observational and field-based studies, data are often highly dependent on the spatial and temporal context of the specific system (Huang 2014, Schnitzer and Carson 2016, LaDeau et al. 2017, Peters and Okin 2017). The weather does not recycle itself. Gradual changes over time, regime shifts, and legacies of past events can greatly influence how natural systems work, as well as our perceptions about how they work (Magnuson 1990). With this complexity, ecology has relied on deep understanding of natural history as a source of ideas about pattern and process (Anderson 2017), often through long periods of intensive observation that could be argued as irreproducible. Even if we could bring back our predecessors, many of their study systems now bear little resemblance to earlier states.

Such irregularities in the natural world can torment experimental ecologists. Repeated or replicated sequences in time are elusive and, likewise, neighboring populations, communities, or ecosystems observed within the same day or year can still differ in important ways that affect the outcomes of studies. Thus, while broad guidelines for reproducible research in the sciences are available (Sandve et al. 2013), ecologists face novel challenges that complicate adoption of such general practices. These challenges are linked to the heterogeneity of the systems we study, as well as the approaches and information we use. Strategies that increase the reproducibility of ecological studies are being pursued (Milcu et al. 2018) and it is important that ecologists address these issues if we are to continue serving a critical role in understanding the complex dynamics of the biosphere in the Anthropocene.