Scalable video coding (SVC) is a new emerging video coding standard to be released as an amendment of the H.264/AVC. The SVC standard adopts four post-processing EC methods to cope with frame loss caused by transmission error. In SVC, however, there have been few studies on burst errors by which several successive frames are corrupted in noisy channels. In this paper, we first propose an error tracking model with which the concealment error and the propagation error can be estimated. Then, in order to cope with the burst error in error-prone networks, we introduce a novel error concealment (EC) strategy based on the proposed error tracking model. For each lost frame, the proposed strategy selects one of four EC methods adopted in the SVC that minimizes the visual degradation.