Incorporating Ecosystem Experiments and Observations into Process Models of Forest Carbon and Water Cycles: Challenges and Solutions

Abstract:

Numerous studies, both experimental and observational, have been conducted over the past two decades in an attempt to understand how water and carbon cycling in terrestrial ecosystems may respond to changes in climatic conditions. These studies have produced a wealth of detailed data on key processes driving these cycles. In parallel, sophisticated models of these processes have been formulated to answer a variety of questions relevant to natural resource management. Recent advances in data assimilation techniques offer exciting new possibilities to combine this wealth of ecosystem data with process models of ecosystem function to improve prediction and quantify associated uncertainty. Using forests of the southeastern United States as our focus, we will specify how fine-scale physiological (e.g. half-hourly sap flux) can be scaled up with quantified error for use in models of stand growth and hydrology. This approach represents an opportunity to leverage current and past research from experiments including throughfall displacement × fertilization (PINEMAP), irrigation × fertilization (SETRES), elevated CO_­2­ (Duke and ORNL FACE) and a variety of observational studies in both conifer and hardwood forests throughout the region, using a common platform for data assimilation and prediction. As part of this discussion, we will address variation in dominant species, stand structure, site age, management practices, soils and climate that represent both challenges to the development of a common analytical approach and opportunities to address questions of interest to policy makers and natural resource managers.