Estimation of Zooplankton Grazing Preferences with Ensemble-based Kalman Filters: Assimilation of CCI Ocean Colour Data
Simon, Ehouarn; Samuelsen, Annette; Bertino, Laurent
Nansen Environmental and Remote Sensing Center, NORWAY

Ocean biogeochemistry models are sensitive to numerous, poorly known parameters. We consider the estimation of positive sum-to-one constrained zooplankton grazing preferences which are introduced to model the relative diet composition of zooplankton species. The sum-to-one constraint cannot be guaranteed by ensemble-based Kalman filters when parameters are bounded. In a previous study (Simon et al., 2012), we suggested variables transformations for the estimation of such constrained parameters and assessed the performances of these approaches within the framework of twin experiments (synthetic observations).

Following these works, we present the results of experiments assimilating real observations. Satellite-derived chlorophyll-a concentrations provided by the ESA CCI Ocean Colour project are assimilated with a Gaussian anamorphosis extension of the deterministic ensemble Kalman filter in the 1D coupled model GOTM-NORWECOM during the period 2003-2004. The configuration is meant to be representative of yearly phytoplankton blooms at station Mike (66°N, 2°W). Because the ESA CCI Ocean Colour data provide estimations of the uncertainties in the observations, e.g. the bias and error variance, we also investigate their impact on the estimation process.

Simon E., Samuelsen A., Bertino L. and Dumont D.: Estimation of positive sum-to-one constrained zooplankton grazing preferences with the DEnKF: a twin experiment, Ocean Sci., 8, 587-602, 2012.