Premium
Assimilation of seasonal chlorophyll and nutrient data into an adjoint three‐dimensional ocean carbon cycle model: Sensitivity analysis and ecosystem parameter optimization
Author(s) -
Tjiputra Jerry F.,
Polzin Dierk,
Winguth Arne M. E.
Publication year - 2007
Publication title -
global biogeochemical cycles
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.512
H-Index - 187
eISSN - 1944-9224
pISSN - 0886-6236
DOI - 10.1029/2006gb002745
Subject(s) - seawifs , environmental science , phytoplankton , photic zone , zooplankton , carbon cycle , biogeochemical cycle , data assimilation , new production , ecosystem model , ecosystem , atmospheric sciences , chlorophyll a , nutrient , oceanography , ecology , biology , geology , meteorology , botany , physics
An adjoint method is applied to a three‐dimensional global ocean biogeochemical cycle model to optimize the ecosystem parameters on the basis of SeaWiFS surface chlorophyll observation. We showed with identical twin experiments that the model simulated chlorophyll concentration is sensitive to perturbation of phytoplankton and zooplankton exudation, herbivore egestion as fecal pellets, zooplankton grazing, and the assimilation efficiency parameters. The assimilation of SeaWiFS chlorophyll data significantly improved the prediction of chlorophyll concentration, especially in the high‐latitude regions. Experiments that considered regional variations of parameters yielded a high seasonal variance of ecosystem parameters in the high latitudes, but a low variance in the tropical regions. These experiments indicate that the adjoint model is, despite the many uncertainties, generally capable to optimize sensitive parameters and carbon fluxes in the euphotic zone. The best fit regional parameters predict a global net primary production of 36 Pg C yr −1 , which lies within the range suggested by Antoine et al. (1996). Additional constraints of nutrient data from the World Ocean Atlas showed further reduction in the model‐data misfit and that assimilation with extensive data sets is necessary.