An innovative way to take the large-scale circulation influence into account in coastal primitive-equation models is explored by an inverse modelling approach. Restricted to barotropic external forcing, this work is a first step in the development of a four-dimensional variational (4DVAR) data-assimilation approach to estimate the best initial and open-boundary conditions that force a coastal model according to interior observations. This development is founded on the OPA modelling system which representation of barotropic coastal dynamics is restricted to motions of long time scales ( a day) due to its rigid lid approximation. Twin experiments are performed in an academic configuration of the Gulf of Lions (located in the northwestern Mediterranean Sea) to study the sensitivity of a remote barotropic forcing to different observational networks measuring surface currents deployed in this area. Three monitoring designs are tested for a large-scale barotropic perturbation in the hindcast mode. It is shown that the space and time distribution of observations acts on the efficiency of the 4DVAR method and then allows coarser datasets.