Any traffic signal control mode requires generation of multiple traffic scenarios for design purposes as well as for validation of the control system. The overall system performance depends on the traffic scenarios used to design the entire system. Therefore, all traffic patterns that might exist in a control system should be considered according to their probability of occurrence on a daily basis. However, when it comes to large networks, considering all combinations of traffic movement levels becomes time-consuming and impractical. A new approach to generate traffic scenarios for large networks is proposed. This approach is based on the selection of significant critical movements controlling the network. Selection of these critical movements is performed by using statistical correlation analysis of actual detector data and synthetic origin-destination analysis for the entire network. The proposed approach was implemented in the design of a traffic-responsive control mode for the Reston Parkway arterial network, in Northern Virginia, which has 14 intersections. Detector data were then used to validate the results of the proposed approach. The validation shows no significant difference between the actual detector data and the proposed procedure's results and therefore proves that the traffic system was correctly modeled and sufficiently represented with the proposed approach.