Big Data-Model Integration
We propose that recent advances in technologies and an emphasis on data and metadata standards and sharing within and across disciplines have positioned ecologists to reduce the probability of future ecological surprises. We developed an operational transdisciplinary approach that accommodates and facilitates integration of large and diverse types of data and knowledge to (a) reduce the high spatial heterogeneity in sampling frequency, intensity, and quality across the surface of the Earth and fill data or knowledge gaps for underrepresented locations; (b) characterize the non-stationarity of environmental drivers and ascertain the extent to which knowledge of the past can or cannot inform the future; and (c) inform land managers and others in prioritizing locations (Peters et al. 2018).
This eight-step Big Data-Model Integration (BDMI) approach with human-guided machine learning and multiple lines of evidence addresses three common phenomena: alternative states, global change, and cross-scale interactions. The key to the BDMI approach is (a) site-based knowledge, (b) expertise in multiple disciplines, and (c) a systems perspective that enables an understanding and quantification of cross-scale interactions through the development of pattern–process relationships at multiple, interacting scales and levels of biological organization.
Key Papers