Configuring an ecosystem model using data from the Bermuda Atlantic Time Series (BATS)

Abstract

Abstract The results of an assimilative approach to guide the configuration of an ecosystem model for the mixed layer of an oligotrophic environment are presented. The time series data from the US Joint Global Ocean Flux Study (JGOFS) Bermuda Atlantic Time Series (BATS), in conjunction with a data assimilation scheme, were used to estimate the model parameters and modify the Fasham et al. (J. Mar. Res. 48 (1990) 591–639) (FDM) model. The evolution of the model from initial to final configuration was driven by: (a) the comparison of the time series data to the model results; (b) analysis of the estimated parameters; (c) observations of the BATS ecosystem from the literature; and (d) corrections of the model pathways. The data assimilation technique was crucial to estimate the optimal parameter set for each of the tested model configurations. The model presented in this paper includes several critical modifications to the FDM model. First, a variable chlorophyll-to-nitrogen ratio is introduced by solving a full equation for chlorophyll a. Second, zooplankton are split into two functional groups: nano/microzooplankton and mesozooplankton. Third, a new formulation is introduced for the microbial loop that is capable of resolving and simulating many of the processes observed in natural environments as well as in laboratory experiments, but had, until now, not been combined in a model. These modifications lead to solving an equation for the temporal evolution of the bio-active dissolved organic-carbon pool. This modified model, in conjunction with data assimilation, allowed us to estimate the model parameters and replicate the annual nitrogen cycle in the upper mixed layer at BATS. Bacteria were found to be a key player in controlling the size of the dissolved organic matter pool and in the amount of regenerated production.