COS 36-5
Evaluation of model size, topology, and currency in systems analysis: comparative network environ analysis of carbon and nitrogen model time series for the Neuse River estuary, USA

Tuesday, August 6, 2013: 2:30 PM
101J, Minneapolis Convention Center
Stuart J. Whipple, Odum School of Ecology and Faculty of Engineering, University of Georgia, Athens, GA
Bernard C. Patten, Odum School of Ecology and Faculty of Engineering, University of Georgia, Athens, GA
Stuart R. Borrett, Biology and Marine Biology, University of North Carolina Wilmington, Wilmington, NC
Background/Question/Methods

An enduring question for network analysis of ecosystem models for holistic ecosystem function is, to what extent do different model size, topology, and currency affect the results and conclusions of such studies? To address this question comparative network environ analysis (CNEA) was conducted on two ecosystem model time series.  The first is a set of sixteen seasonal steady-state network models of nitrogen (N) storage and flow in the Neuse River estuary (NRE) constructed by R. R. Christian and colleagues.  The second is a set of four network models of carbon (C) storage and flow in the NRE constructed by R.R Christian, D. Baird and colleagues. Network environ analysis (NEA) provides quantitative measures of direct and indirect relationships between compartments and their environments within defined systems. Each environ's dynamics are unique, and they sum over all the environs to give the aggregate dynamics of the whole as measurable by empirical methods.  By comparing measures of environ autonomy for the larger C-based NRE models to the smaller N-based NRE models, we will evaluate the degree to which the size, topology, and model currency (C or N) in the Neuse River models serve to determine CNEA results.

Results/Conclusions

Perhaps the most important implication of this presentation is the potential for CNEA to provide a way compare models of the same ecosystem with different currencies to determine how they process inputs across their boundaries. Despite differences in aggregation level (7 vs 30 compartments), and topology between these N and C models, the compartments types included and turnover rate among comparable compartments show very similar patterns. These similarities provide a basis for comparison of the models. It was found that large inputs to nutrient, algal, and sediment stocks in realized environs resulted in different storage responses by the N and C models.  In the N model, large storages appear in the dissolved nutrient compartments (NO3 /NO2, DON, NH4), while in the C model large storages appear in the suspended and sediment POC compartments.  These results have implications for the use of model analysis results; different model currencies, and topology, especially the degree of biogeochemical cycling, may result in very different conclusions regarding ecosystem biogeochemical processing.