The Next Decade of Big Data in Ecosystem Science

Abstract

Ecosystem scientists will increasingly be called on to inform forecasts and define uncertainty about how changing planet conditions affect human well-being. We should be prepared to leverage the best tools available, including big data. Use of the term ‘big data’ implies an approach that includes capacity to aggregate, search, cross-reference, and mine large volumes of data to generate new understanding that can inform decision-making about emergent properties of complex systems. Although big-data approaches are not a panacea, there are large-scale environmental questions for which big data are well suited, even necessary. Ecosystems are complex biophysical systems that are not easily defined by any one data type, location, or time. Understanding complex ecosystem properties is data intensive along axes of volume (size of data), velocity (frequency of data), and variety (diversity of data types). Ecosystem scientists have employed impressive technology for generating high-frequency, large-volume data streams. Yet important challenges remain in both theoretical and infrastructural development to support visualization and analysis of large and diverse data. The way forward includes greater support for network science approaches, and for development of big-data infrastructure that includes capacity for visualization and analysis of integrated data products. Likewise, a new paradigm of cross-disciplinary training and professional evaluation is needed to increase the human capital to fully exploit big-data analytics in a way that is sustainable and adaptable to emerging disciplinary needs.