Retaining consistency for large knowledge bases is a difficult task. This holds especially in the case where the knowledge base comprise temporal knowledge and where the knowledge comes from independent and unreliable sources. In this paper we propose the use of temporal logics, i.e., CTL, to describe the background theory and the corresponding Kripke Structure to store the temporal knowledge. Moreover, we introduce a declarative formalization of belief revision which is necessary to keep the knowledge base in a consistent state. Finally, we discuss how the structure of CTL formulas can be used to implement belief revision. The research described in the paper is motivated by a project that deals with automating the analysis of meetings, e.g., to provide meeting summaries, where cameras, microphones, and other sources of knowledge has to be integrated.