An Ecosystem Model Comparison on the Northeast U.S. Continental Shelf Using Data Assimilation

Yongjin Xiao, Virginia Insitute of Marine Science, yxiao@vims.edu (Presenter)
Marjorie Friedrichs, Virginia Institute of Marine Science, marjy@vims.edu

The last two decades have witnessed significant advances in high-resolution numerical models of the coastal ocean carbon cycle. While the large-scale hydrodynamics can now be reasonably well reproduced, the ecosystem counterpart remains poorly developed. Recent efforts have been directed toward incorporating complex food webs, including multiple phytoplankton and zooplankton compartments. This naturally leads to the question: how many phytoplankton (P) and zooplankton (Z) compartments should be included in these models in order to accurately simulate coastal carbon cycles? Here we address this question by implementing five ecosystem model variants in a 1D assimilative (variational adjoint) model testbed at multiple sites along the northeastern U.S. continental shelf. The five models resemble one another except for variations in the level of complexity included in the lower trophic levels, which range from a simple 1P1Z food web to a considerably more complex 3P2Z food web. The assimilation of satellite-derived chlorophyll, particulate organic carbon and productivity fields allow for a rigorous comparison of the skill of the model variants. Pre-assimilation simulations reveal significant inter-model differences in surface bloom magnitude and timing, with the largest differences found closest to shore. The most significant differences also occurred between the 1P1Z model and the 2P1Z model, with additional increases in complexity resulting in smaller differences in simulated fields. Rigorous data assimilative comparisons indicated that the 1P1Z model generated reasonable model-data misfits at individual sites, but the parameters optimized for those individual sites produced large misfits when used at other sites along the U.S. east coast. When assimilating data from multiple sites, the 2P models showed the lowest model-data misfits whereas the 2Z models did not show substantial improvement compared to the 1Z models. These preliminary results suggest that incorporating a third P or second Z may not be required for this study region.

Presentation: 2011_Poster_Xiao_294_72.pdf (1034k)

Presentation Type:  Poster

Session:  Global Change Impact & Vulnerability   (Tue 11:30 AM)

Associated Project(s): 

• Friedrichs, Marjy: The Impacts of Climate Change on Phytoplankton Community Structure along the U.S. Northeastern Continental Margin ...details

Poster Location ID: 294