Volumes and Issues  Contents of Issue 3  Special Issue  
Ocean Sci. Discuss., 3, 503-539, 2006
www.ocean-sci-discuss.net/3/503/2006/
© Author(s) 2006. This work is licensed
under a Creative Commons License.


Reduced-order optimal interpolation for biomass assimilation

G. Crispi, M. Pacciaroni, and D. Viezzoli
Istituto Nazionale di Oceanografia e di Geofisica Sperimentale, Italy

Abstract. The analysis of the evolution of the chemical and biological characteristics of the Mediterranean marine ecosystem requires an integrated approach. Consistently, an ecosystem description, embedded in the MFSTEP one eighth degree three-dimensional general circulation model, is tested and used following a twin experiment approach. The ecosystem model is based on the NPZD trophic chain: inorganic nitrogen, N, phytoplankton, P, zooplankton, Z and detritus, D. Assimilation of synthetic biomass data is performed by means of the reduced-order optimal interpolation system SOFA.

The synthetic "sea-truth" data are daily averages obtained from a sixty-nine days reference run (RR). The twin experiments consist in performing two runs: a free run (FR) with wrong summer-depleted phytoplankton initial conditions and an assimilated run (AR), in which, starting from the same FR wrong phytoplankton concentrations, weekly averaged surface biomasses extracted from the RR results are assimilated.

The comparison of the FR results with the AR ones shows a good convergence, on a basin Mediterranean scale, confirming improvements of the forecasting in each of the four ecological compartments.

Regional trophic regimes are analysed and interpreted in the western and eastern Mediterranean subbasins, for explaining the deteriorating behaviour of the total nitrogen.


Discussion Paper (PDF, 2774 KB)   Interactive Discussion (Closed, 4 Comments)   Final Revised Paper (OS)   

Citation: Crispi, G., Pacciaroni, M., and Viezzoli, D.: Reduced-order optimal interpolation for biomass assimilation, Ocean Sci. Discuss., 3, 503-539, 2006.   Bibtex   EndNote   Reference Manager

Recent Papers