The PEcAn Project: Carbon-Cycle Reanalysis Facilitated by Model-Data Ecoinformatics

The fundamental questions about how terrestrial ecosystems will respond to climate change are straightforward and well known, yet a small number of important gaps separate the information we have gathered from the understanding required to inform policy and management. A critical gap is that no one data source provides a complete picture of the terrestrial biosphere, and therefore multiple data sources must be integrated in a sensible manner. Process-based models represent an ideal framework for this synthesis, but to date model-data synthesize has only made use of a subset of the available data types, and remains inaccessible to much of the scientific community, largely due to the daunting ecoinformatics challenges. The Predictive Ecosystem Analyzer (PEcAn) is an open-source scientific workflow system and ecoinformatics toolbox that manages the flow of information in and out of regional-scale terrestrial biosphere models, facilitates formal data assimilation, and enables more effective feedbacks between models and field research. PEcAn makes complex analyses transparent, repeatable, and accessible to a diverse array of researchers. PEcAn is not model specific, but rather encapsulates any ecosystem model within a set of standardized input and output modules. Herein we demonstrate PEcAn's ability to automate many of the the tasks involved in modeling by gathering and processing a diverse arrays of data sets, initiating ensembles of model runs, visualizing output, and comparing models to observations. PEcAn employs a fully Bayesian approach to model parameterization and the estimation of ecosystem pools and fluxes that allows a straightforward propagation of uncertainties into analyses and forecasts. This approach also makes possible the synthesis of a diverse array of data types operating at different spatial and temporal scales and to easily update predictions as new information becomes available. We also demonstrate PEcAn's ability to iteratively synthesize information for literature trait databases, ground observations, eddy-covariance towers, and remote sensing and to quantify the reductions in overall uncertainty as each new dataset is added. PEcAn also automates a number of model analyses, such as sensitivity analyses, ensemble prediction, and variance decomposition which collectively allow the system to partition and ascribed uncertainties to different model parameters and processes. PEcAn provides a direct feedback to field research by further automating the estimation of sample sizes and sampling distributions required to reduce model uncertainties, enabling further measurements to be targeted and optimized. Finally, we will present the PEcAn development plan and timeline, including new features such as the synthesis of remotely sensed data, regional-scale data assimilation, and real-time forecasting. Ultimately, PEcAn aims to make ecosystem modeling and data assimilation routine tools for answering scientific questions and informing policy and management.