Development of a Data Assimilation Capability Towards Ecological Forecasting in a Data-Rich Era

点击翻译按钮获取中文摘要

The University of Oklahoma is awarded a grant to develop an Ecological Platform for Assimilation of Data (EcoPAD) for data assimilation and forecasting in ecology. EcoPAD will include components of (1) core computational algorithms (e.g., ecological models) that are specifically designed to solve ecological issues, (2) a variety of optimization techniques for data assimilation, (3) various data bases that will feed into EcoPAD, and (4) diverse functions of EcoPAD. The functions enable users to (i) estimate model parameters or state variables, (ii) quantify uncertainty of estimated parameters and projected states of ecosystems (e.g., carbon sinks in USA), (iii) evaluate model structures, (iv) assess sampling strategies, and (v) conduct ecological forecasting. EcoPAD will be developed, tested, and applied at three spatial scales. First, EcoPAD will be developed against plot-level data from experiments in the Duke Forest and Oklahoma prairies to examine its capability of extract information on ecosystem responses to climate change. Second, the project will develop EcoPAD at the AmeriFlux network of eddy-flux measurements in the North America to examine its capability of forecasting diurnal, seasonal, and interannual dynamics of net ecosystem exchange. Third, EcoPAD's capability of forecasting future states of ecosystems and their services will be tested and evaluated at regional and continental scales. EcoPAD has the potential to become a powerful eco-informatics tool that assimilate data from measurement sensor networks and to generate data products that will be useful for policy making on resource management and climate change mitigation. This project will have integrated teaching, training, and learning through student and post-doc training, integrated courses in ecological modeling, and outreach to K-12 students and teachers. NEON and other sensor networks have been designed to provide better ecological service to the society. Data assimilation and ecological forecasting are key components for success of those networks. This project will develop EcoPAD to extract information from massive data for ecological forecasts, which can directly serve the society and be useful for policy making. Results from this project will be useful for education of public on ecosystem services. For further information see the PI's lab website at http://bomi.ou.edu/luo/.