Search

Monitoring technologies - from satellites to smartphones- are creating information about global ecosystems at a rate and resolution that was unfathomable even a decade ago. Coupled with advances in computational tools (e.g., computer vision), these billions of observations are readily translated into usable derivative data products and made available on data-sharing platforms (e.g., GBIF, Movebank) (Wuest et al., 2019), promising unprecedented insight into macroecological processes and decision support for a more sustainable planetary future. These data increasingly act as a shaping social force: used by governments to designate conservation priorities, by fintech companies to price biodiversity offsets and help comply with pending legislation, and by machine learning experts to benchmark and apply new algorithmic tools to real-world data (Luccioni and Rolnick., 2022). However, widespread use of these ecosystem observations often belies a reality that the species these data tell us most about is the one species they never intended to include -- humans. We see not only roads, cities, and the rise of monitoring technology reflected in the billions of biodiversity observations (Hughs et al., 2021), but shadows of a colonial past (Zizka et al., 2020), the weekly sway of work schedules in our contemporary capitalist society (Żmihorski et al. 2012), and echoes of our racial and economic disparities (Ellis-soto et al., 2022).With urgent calls to develop biodiversity metrics that hold countries, communities, and companies accountable for their commitments to the post-2020 Kunming-Montreal global biodiversity framework (GBF), understanding the histories and human dimensions of biodiversity data is critical to ensuring policy and practice informed by these data don’t exacerbate past and present inequities. This work explores how uncorrected socio-political disparities in underlying biodiversity data impact not only our insights about ecosystem processes, but the distributional equity of decisions derived from those data. We explore how careful statistical models can help identify and control for social and political data disparities - a start at disentangling the observer from the observed - but only to the extent that we can identify and quantify those disparities. Moreover, we show how the feedbacks between data disparities and decision biases in the environmental domain are complex. Understanding the new generation of global environmental data, particularly data derived from participatory platforms, requires expertise in social, cultural, and political processes underlying these data infrastructures and histories, just as much as it requires more complex statistical methods and ecological knowledge. We address how appropriately dealing with the new era of ecological data requires the ability to collaboratively leverage local knowledge in global analyses and borrow strength across different data types and scales.