Modelling ecosystem functions and properties at different time and spatial scales in shallow coastal lagoons:: An application of the LOICZ biogeochemical model

The Land-Ocean Interactions in the Coastal Zone (LOICZ) biogeochemical model (LBM) was applied at different temporal and spatial scales in 17 Italian lagoons of the LaguNet network (http://www.dsa.unipr.it/lagunet). A series of alternative assumptions taking into account benthic vegetation and sedimentary fluxes were introduced and compared with the classical LBM approach at various time scales. The reliability of the LBM application to the seventeen Italian lagoons was tested by comparison to a pool of shallow coastal systems from the global LOICZ database with comparable depths and sizes. The nutrient loads of the Italian sites can be considered relatively low, particularly for dissolved inorganic phosphorus (DIP). Although the median values of estimated internal transformations (source-sink) of both dissolved inorganic phosphorous and nitrogen at the LaguNet sites were comparable with the selected LOICZ sites, the positive and negative extreme values were one order of magnitude lower. Overall, the LBM applications to the Italian sites gave good quality budgets for shallow systems subjected to relatively low nutrient inputs and with a wide range of primary producer communities, including seagrass, macroalgae and phytoplankton. Furthermore, stoichiometry of Carbon: Nitrogen: Phosphorous for the different primary producer groups allowed the integration of previous studies by identifying a series of relationships between nutrient loads and ecosystem functions. To some extent, the LBM application at monthly and seasonal time scales, as shown for the Sacca di Goro lagoon, can simulate the internal variability affected by the life cycle of the dominant primary producers. Therefore, the LBM not only allows the assessment of net ecological metabolism but seems, also, capable of representing the wide range of trophic conditions associated with shallow coastal systems. Overall, the application of alternative assumptions supports the robustness of the classical LBM, although some of the simplifications that enable the LBM to function in a wide range of systems and with incomplete data sets must be considered with caution for shallow environments. (c) 2007 Elsevier Ltd. All rights reserved.