Our goal is to develop an operational trans-disciplinary approach to complex landscapes that facilitates integration of large and diverse types of data and knowledge to: (1) reduce the high spatial heterogeneity in sampling across heterogeneous landscapes and fill data/knowledge gaps for underrepresented locations; (2) characterize the non-stationarity of environmental drivers and ascertain the extent to which knowledge of the past can, or cannot, inform the future; and (3) inform land managers and others in prioritizing locations. We illustrate the utility of our approach to extend process-based understanding of ANPP at intensively studied locations to dynamics of ANPP at other under-represented locations, and to predict spatial variability in ANPP across the landscape.