SUMMARY The success of the marine controlled-source electromagnetic (CSEM) method as an exploration tool for evaluating the in situ resistivity of targets, prior to drilling, logically leads to the suggestion that the method could be used to monitor reservoirs during production. Simple model studies suggest that under favorable circumstances the effect of draining a reservoir produces an approximately linear response in CSEM amplitudes (that is, a 1% change in reservoir volume produces about a 1% change in electric fields). The vertical electric field generates the strongest response, and the response is spatially correlated with the location of reservoir depletion. Although ‘geological noise’ and the inherent non-uniqueness of geophysical interpretation would make a 1% target response dubious for exploration purposes, since we can infer that the only source of changes in CSEM response are due to changes in the reservoir, the method is appealing for development applications. The only difficulty is that to avoid corrupting the signal associated with changes in the electric and magnetic fields, the survey geometry needs to be maintained to a fraction of a percent. While this may be challenging using current exploration CSEM technology, we propose to use ROV (re-)deployment of receivers, and perhaps transmitters, in order to obtain reproducible geometries. INTRODUCTION Use of the marine controlled-source electromagnetic (CSEM) method for identifying hydrocarbon reservoirs was first demonstrated seven years ago in a test by Statoil over the Girassol prospect, offshore Angola (Ellingsrud et al., 2002). Since that test, the concept has been embraced with some enthusiasm by the exploration industry, mainly as a tool for assessing the resistivity of targets identified by seismic surveys prior to drilling (e.g. Constable and Srnka, 2007). To a lesser extent, it has been used for estimating the size and extent of reservoirs and as a reconnaissance tool. The logical next step is to take the methodology and move it from the exploration environment to the production environment, as a means to monitor the geometry of a reservoir as it is developed. The purpose of this paper is to present a simple feasibility study of this so-called 4D-CSEM approach as a prelude to carrying out a field study. 1D ANALYSIS As is often the case, it is simple yet instructive to start with a 1D experimental design study, and when the source and receiver are both over a 3D tabular reservoir, the response is well represented by 1D theory (Constable and Weiss, 2006). We used the algorithm and code of Flosadottir and Constable (1996) to study the canonical 1D reservoir, a 100 Wm oil. The biggest effect is for top flooding, since this part of the model is closer to the seafloor, suggesting that larger flooding from the bottom may be confused with top flooding. Indeed, at a range of 4–5 km the amplitude and phases are similar as for 10% top flooding, but the behavior at shorter and longer ranges is markedly different, making it easy to tell the two scenarios apart.