PalEON-a paleocological observatory network to assess terrestrial ecosystem models
- Funded By: National Science Foundation
- ECI Investigators: Jason McLachlan, Melissa Berke, Alan Hamlet, Adrian Rocha
- Partners: Boston University, Harvard Forest, Michigan State University, University of Arizona, University of California, University of Colorado, University of Montana, University of Wisconsin, West Virginia University, Yale University
Sometimes scientists can get a better grasp of the future by digging deeper into the past. The Paleoecological Observatory Network (PALEON) is a team of over 65 paleoecologists, environmental statisticians, and ecosystem modelers.
Their goal is to collate existing data and collect new historical and paleoecological data from Northeastern and Midwestern states and from the central boreal forest of Alaska. Statistical models are then applied that make inference from these data with uncertainty about changing terrestrial ecosystems over the last 2000 years; and to integrate this statistical inference into models of ecosystem change. PalEON’s ultimate goal is to assimilate inference from long-term data into models, so that the slow processes influencing projections of ecosystem change are constrained by data.
PalEON continues to develop tools and datasets for facilitating data-model fusion for long-term historical and paleoecological data. Additionally, PalEON continues to develop Bayesian hierarchical models of long term trends in wildfire; climate; and forest composition, structure, and productivity. These estimates are being used to validate long-runs of a suite of ecosystem models, which will improve inference from these models through data assimilation and model development.
In this project, the research team collects a substantial database of historic and prehistoric forest data. These data range from the surveys used by early European settlers, often collected in dusty town archives, to the fossil remnants of leaves and pollen preserved in freshwater lake sediments. Then the researchers analyze these “paleoecological” data with Bayesian statistical tools to identify robust patterns of past forest change. The lessons learned from this “hindcasting” exercise will be used to improve ecosystem models, allowing for more confident anticipation and planning of the future.