Vortices, the regions of swirling coherent motion of fluid, are of fundamental importance in understanding the dynamics of turbulent flows. Recent advances in computational and experimental resources have resulted in massive volumes of highly resolved flow field data. Identification of coherent vortex structures from these space-time discretized flow dataset is the key issue of vortex identification. We consider identification schemes based on pointwise analysis of the velocity gradient tensor. A new measure of the local spatial coherence in a vortex is introduced. Different criteria are compared for two classes of turbulent flows: Isotropic and sphere wake. Remarkably similar vortex structures are observed using the Q, λ2 and swirling strength criterion. An explanation based on swirling strength and the proposed local coherence measure is offered for this observation.