In this paper a novel Observer-driven Reduced Copy (ORC) approach is proposed to deal with communication network data-dropouts in a smart power grid with large-scale deployment of distributed and networked Phasor Measurement Units (PMUs) and wind energy resources, which uses knowledge of the nominal system dynamics during data dropouts to improve the damping performance under large disturbances, where the conventional feedback would suffer. To that end a reduced order 16-machine 5-area dynamic equivalent model of the New England-New York power system with replacement of one existing synchronous generator and a power system stabilizer (PSS) by a DFIG-based wind farm (WF) is considered. The problem with electromechanical oscillation damping control through WFs using locally available signals is identified and a systematic approach for selection of control input and remote feedback signals through modal analysis is presented. The remote feedback signals sent through communication channels encounter data dropout which is represented by the Gilbert-Elliott model. Moreover, an expression for the bound on the error norm between the actual and the estimated states relating data dropout and model mismatch is also derived. Nonlinear time-domain simulations demonstrate that the ORC gives significantly better performance compared to the conventional feedback under higher data drop situations.