A material balancing scheme for ocean colour data assimilation

Abstract

A material balancing scheme for assimilation of ocean colour data in the FOAM-HadOCC model (Forecasting Ocean Assimilation Model with Hadley Centre Ocean Carbon Cycle Model biogeochemistry) has been developed with the aim of exploiting satellite data to improve air-sea CO2 flux estimates. The balancing scheme uses surface chlorophyll increments to determine increments for the biogeochemical tracers: nutrient, phytoplankton, zooplankton, detritus, dissolved inorganic carbon (DIC) and alkalinity.



The analysis conserves carbon at each grid point and nitrogen at grid points where sufficient nitrogen is available, on the assumption that the modelling of processes transferring material between biogeochemical compartments is the main source of error. Phytoplankton increments are calculated using the model nitrogen:chlorophyll ratio. Increments to the other nitrogen pools (nutrient, zooplankton and detritus) are determined by a balancing model that responds to changes in the plankton dynamics. The nutrient balancing factor, the fraction of the phytoplankton increment to be balanced by the nutrient increment, varies according to the relative contributions of growth and loss rate errors to the phytoplankton error, as estimated from a probability model. Preliminary balancing factor values are adjusted to satisfy state-dependent restrictions on the size of the increments. Increments derived in this way are applied down to the depth of the mixed layer. Further increments are applied where necessary to avoid the creation of unrealistic sub-surface nutrient minima. Increments to DIC balance the implied carbon changes in the organic compartments and alkalinity increments are inferred from those for nutrient.



An off-line evaluation of the scheme is carried out in a 1-D test-bed in which HadOCC biogeochemistry is forced by physical data for a range of latitudes in the eastern North Atlantic. Evaluation is by twin experiments for which synthetic system trajectories are generated by perturbing model parameters during integration to provide a range of plausible truths. Assimilation of daily chlorophyll observations, with or without simulated observation error, gives major improvements in pCO2 at the high latitudes but less improvement at low latitudes where it has a detrimental effect on summer and early autumn pCO2 due to errors in the model nitrogen:chlorophyll ratio. Beneficial effects of nitrogen balancing are demonstrated by comparison with experiments in which only phytoplankton and DIC are updated. The sub-surface nutrient correction increments are shown to reduce, but not remove, undesirable effects

of assimilation on the nutrient and phytoplankton profiles.

More information

Identifiers

Catalogue record

Export record