The relative importance of a node in wireless mesh networks (WMNs) is typically related to its topological position and data forwarding capacity. In dynamic scenarios, however, these particular characteristics may change over time. In the analysis of complex networks, the concept of network centrality reflects the relative importance of a node within a network. We first show the current centrality metrics are unable to dynamically classify nodes in the scenarios we target. We then propose the Dynamic Labeling of Nodes (DyLaN), a new metric that aims at identifying and monitoring the most important nodes in time-varying WMN scenarios. We evaluate DyLaN using real-world traces showing that it can improve the quality of managing such dynamic WMNs.