Abstract: The inverse problem of seawater intrusion(SWI) is reviewed. It represents a challenge because of both conceptual and computational difficulties and because coastal aquifer models display many singularities:(1) head measurements need to be complemented with density information; (2) salinity concentration data are
very sensitive to flow within the borehole. Data problems can be reduced by incorporating the measurement process within model calibration; (3) SWI models are extremely sensitive to aquifer bottom topography; (4) the initial conditions may be far from steady state and depend on the location and type of sea-aquifer connection. Problems with aquifer geometry and initial conditions can be addressed by parameterization, which allows for modification during inversion. The four sets of difficulties can be partly overcome by using tidal response and electrical conductivity data, which are highly informative and
provide extensive coverage. Still, SWI inversion is extremely demanding from a computation point of view. Computational improvements are discussed.