B11C-0452
What measurements are needed to capture coupled carbon and nitrogen cycles in arctic tundra?

Monday, 14 December 2015
Poster Hall (Moscone South)
Kelseyann Kremers and Adrian V Rocha, University of Notre Dame, Notre Dame, IN, United States
Abstract:
Greening has been observed across the arctic, but the ecological processes that enable widespread increases in plant productivity have been difficult to understand with field measurements alone. Using the functional convergence of foliar nitrogen and leaf area index in arctic plants, we developed a simple coupled carbon and nitrogen cycling model (CCaN) to increase our understanding of warming induced vegetation changes and carbon and nitrogen coupling in the arctic. We used primary literature and data from long term ecological research sites at Toolik Field Station, Alaska to calculate prior ranges for CCaN parameters, and then we assimilated five years of eddy covariance data from a nearby field site using a batch data assimilation method based on Markov Chain Monte Carlo techniques. Variance decomposition analyses show that the majority of model variance in carbon and nitrogen stocks can be attributed to uncertainty in four parameters: proportion of nitrogen in foliage, proportion of nitrogen in roots, nitrogen uptake rate, and litter rate. These parameters associated with nitrogen cycling are widely used in biogeochemical cycling models and are difficult to constrain because they vary greatly across plant functional groups. The parameters responsible for variance in net ecosystem exchange vary seasonally; winter variance is controlled by parameters associated with the temperature sensitivity of heterotrophic respiration, and summer variance is controlled by the proportion of nitrogen in foliage. The widespread greening observed across the arctic over the last decade has been attributed to the direct effects of increased temperature, despite the inconsistency of the response of arctic plants to experimental warming indicating there must be other mechanisms at play. These mechanisms can be teased apart using models, but we must first improve model predictions by constraining widely-used processes and parameters, particularly those linked to nitrogen cycling.