|Hunt, Earle - Ray|
Submitted to: Ecological Monographs
Publication Type: Peer reviewed journal
Publication Acceptance Date: 11/24/2003
Publication Date: 8/30/2004
Citation: Hanson, P.J., Amthor, J.S., Wullschleger, S.D., Wilson, K.B., Grant, R.F., Hartley, A., Hui, D., Hunt, E.R., Johnson, D.W., Kimball, J.S., King, A.W., Luo, Y., McNulty, S.G., Sun, G., Thornton, P.E., Wang, S.S., Williams, M., Cushman, R.M. 2004. Carbon and water cycle simulations for an upland oak forest using 13 stand-level models: Intermodel comparisons and evaluations against independent measurements. Ecological Monographs. 74(3):443-489. Interpretive Summary: Computer simulation models are designed to predict ecosystem carbon and water cycles for different purposes, from integration of experimental results to management applications. The model designer makes important decisions on spatial scale, temporal scale, and physiological complexity to best fulfill the model's purpose. Models are tested with independent data and evaluated for agreement with the observations. If the predictions and observations do not agree, then if necessary the model is modified, and evaluations with additional data are performed. If the model is useful, it is often adopted for other purposes, which requires additional modifications. Oak Ridge National Laboratory conducted an experimental manipulation of the water supply to an oak-dominated ecosystem, called the Walker Branch Throughfall Displacement Experiment, and used the experimental observations from the years 1993 to 2000 to test and compare 13 different ecosystem models. The 13 models ranged in complexity from detailed models specific to oak forests to general models that can be applied from deserts to rain forests. The time steps for the 13 models ranged from hourly, to daily, to annually. There was not one best model for all of the time steps for all of the variables measured in the experiment. The mean output from the different models for a single variable was almost always a better match to the data than the outputs of the 13 individual models. Taken together, detailed models and hourly time steps were better than general models and daily time steps. All models did not do as well under drought as under normal climatic conditions, so more work in model development is necessary. For questions concerning possible ecosystem responses to global climate change, models may be the best tool available. These results suggest that comparisons of multiple models will provide more reliable predictions of ecosystem responses.
Technical Abstract: Models represent our primary method for integration of small-scale, process-level phenomenon into a comprehensive description of forest-stand or ecosystem function. They also represent a key method for testing hypotheses about the response of forest ecosystems to multiple changing environmental conditions. This paper describes the evaluation of 13 stand-level models varying in their spatial, mechanistic, and temporal complexity for their ability to capture intra- and inter-annual components of the water and carbon cycle for an upland, oak-dominated forest of eastern Tennessee. Comparisons between model simulations and observations were conducted for hourly, daily, and annual time steps. Measured data for the comparisons were obtained from a wide range of methods including: eddy covariance, sapflow, chamber-based soil respiration, biometric estimates of stand-level net primary production and growth, and soil water content by time or frequency domain reflectometry. Response surfaces of carbon and water flux as a function of environmental drivers, and a variety of goodness-of-fit statistics (bias, absolute bias, and model efficiency) were used to judge model performance. A single model did not consistently perform the best at all time steps or for all variables considered. The mean of all model outputs, however, was nearly always the best fit to observations. Not surprisingly, models missing key forest components or processes, such as roots or modeled soil water content, were unable to provide accurate predictions of ecosystem responses to short-term drought phenomenon. Nevertheless, an inability to correctly capture short-term physiological processes under drought was not necessarily an indicator of poor annual water and carbon budget simulations. This is possible because droughts in the subject ecosystem were of short duration and therefore had small cumulative impact. Models using hourly time steps, detailed mechanistic processes, and having a realistic spatial representation of the forest ecosystem provided the best predictions of observed data. General models designed for application to a range of ecosystems had the largest errors in this intercomparison. Predictive ability of all models deteriorated under drought conditions suggesting that further work is needed to evaluate and improve ecosystem model performance under unusual conditions that are the focus of environmental change questions.