Modelling Nutrients and ChlorophyllaTime Series in an Estuarine Upwelling Ecosystem (Ría de Vigo: NW Spain) Using the Box–Jenkins Approach

Hydrographic variables were monitored twice weekly at a fixed station in the Ría de Vigo (NW Spain) from 1987. In this work, the behaviour of nutrients and chlorophyllain response to external forcing has been analysed. To achieve this objective, the time course of nutrients and chlorophyllahas been investigated through the Box–Jenkins approach, by means of both univariate and transfer function–noise (TFN) models. Meteorological and thermohaline variables were used to characterize the external forcing in the building of TFN models. Nutrients in the Ría de Vigo can be grouped into different categories depending on the factors which control their behaviour. Surface NO3−, DIN and Si(OH)4, and bottom NH4+, NO3−, and DIN are strongly influenced by external factors such as runoff and/or upwelling–downwelling events. Surface NH4+and HPO42−are clearly under internal control, i.e. biologically mediated processes operating at lower time-scales which override the larger scale hydrodynamic (or weather-induced) forcing. Bottom HPO42−and Si(OH)4are under both internal and external control. Box–Jenkins methods applied in this work are conceptually too simple to model the complex behaviour of phytoplankton biomass. However, modelling chlorophyllausing this approach may be a valuable tool to focus our attention on situations when large fluctuations of phytoplankton biomass occur