Water exchange of the Stockholm archipelago—a cascade framework modelling approach

The Stockholm archipelago spans roughly a semicircular area with a radius of approximately 60 km, traditionally partitioned into three parts: the inner, the middle and the outer archipelago. This subdivision coincides with differing water exchange regimes. The inner and middle archipelagos are characterised by comparatively larger basins which are interconnected by a limited number of straits. This configuration is well suited for a discrete basin (DB-) model approach by partitioning the area into a set of sub-basins that are only resolved vertically. The advantage of this approach over 3D-models is the possibility for enhanced vertical resolution and improved strait exchange formulation, outweighing the disadvantage of neglected horizontal gradients within the basins. In the inner archipelago the dominating exchange process is estuarine circulation, induced by the marked freshwater discharge and the vertical mixing. In the outer and middle archipelagos the density fluctuations due to Ekman pumping along the Baltic boundary interface produce another type of baroclinic process that clearly dominates. Measurements to adequately resolve these density variations do not exist. Missing forcing data are provided by linking the middle archipelagoʹs boundary straits to a 3D-model of the Baltic with a grid resolution of 0.5 nautical miles (n.m.). This fine resolution model (FR-domain) is in turn driven by the atmospheric forcing and the density variation at the rectangular boundary of the FR-domain which acceptably resolves both the interfacial straits and the outer archipelagoʹs complex hypsography. Massive computing resources would be demanded if the FR-domain were extended to comprise the entire Baltic. The FR-domain is thus interfaced with an existing coarse resolution model of the entire Baltic (CR-domain) with a grid size of 5 n.m., the open boundary of which is located in the Kattegat. This 3-fold model set-up has been run for one whole year (1992) with a one-year spin-up time to make up for the lack of initial data. The model concept is at this stage to be regarded as a framework for further development in anticipation of improved formulations, particularly for the strait exchange formulation. Therefore only primary validation experiments and a few sensitivity analyses have been performed.