Biosphere 2 Science: Large Scale Research with Big Data Needs
The University of Arizona uses the biomes and experimental systems at Biosphere 2 (B2), including the Tropical Rainforest, Ocean, and Landscape Evolution Observatory (LEO), to study ecosystem responses to controlled disturbances. The large scale of this research infrastructure enables controlled ecosystem studies at realistic spatial scales with high observational density—effectively bridging the gap between laboratory and field studies. Observations of the B2 systems is facilitated by dense sensor networks that stream long-term continuous data, for example over 1 million data points are read from nearly 2,000 sensors each day at LEO. These data are integrated with discrete in situ and or lab-derived observations, for example soil chemistry and microbial community composition, to extend the types of interdisciplinary research questions that can be addressed. Even on a routine basis, integrating the interdisciplinary and heterogenous data generated at B2 is a challenge, and this task is even more daunting during large research campaigns in which additional instrumentation, expertise, and collaborators generate massive data sets to describe ecosystem response to controlled experiments over relatively short periods. As an example, the upcoming B2 Water, Atmosphere, and Life Dynamics (B2 WALD) campaign will bring together an international team to constrain the Tropical Rainforest response to drought with continuous in situ datasets (online gas analysis and environmental sensors), with discrete in situ measurements (portable gas and leaf spectral analysis) and discrete lab-based insights (trapped gases, leaf traits, soil & water chemistry, multi-omics datasets). Uniting these data is a significant challenge, but an essential basis to derive the most impactful results from experimentation at B2. In this talk, I will describe current and planned experiments leading to data generation at Biosphere 2 with the goal of stimulating discussions on how current approaches in computer science, mathematics, and statistics may enhance scientific outcomes in this unique facility.