This study investigates the performance of a vibration-based technique for damage assessment of reinforced concrete bridges from non-stationary and incomplete acceleration response measurements during high amplitude earthquakes. The proposed damage assessment technique is targeted to be used in the aftermath of a major earthquake event to rapidly and remotely assess the functionality status of the bridge and identify potential hazards to the public safety. As the first step of the procedure, time-frequency representation of the response of the bridge is achieved by applying stochastic subspace system identification technique to successive and overlapping windows of the response measurements. The timefrequency representation is then used to identify the longest ending segment of the response with relatively stable modal properties. Post-earthquake experimental modal properties of the bridge are subsequently extracted from the identified stable portion of the response. These properties are used to estimate the amount of degradation in stiffness of the structural elements through an optimization-based finite element model updating technique. The Genetic Algorithm optimization technique is used to update the stiffness properties of the structural elements by minimizing the error between analytical and experimental modal properties of the bridge. The proposed damage assessment procedure is applied to experimental data from a large-scale shake table test during which a quarter-scale model of a reinforced concrete bridge was subjected to a series of earthquake and low-amplitude white noise base excitations. The meaningful agreement between the stiffness correction factors identified from both types of motions at the same damage state of the bridge demonstrates that the proposed procedure can effectively be applied for post-earthquake damage assessment of the bridges from nonlinear responses during high amplitude earthquakes.© (2011) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.