We study the resilience of interdependent critical infrastructure networks where the state of the two networks depend on each other. Specifically, this research focuses on a water delivery system and the power grid where the water network requires electricity from the power grid and the power stations rely on clean water to produce electricity. The dynamic nature of interdependent links between water and power infrastructure systems leads to challenges in evaluating the performance and assessing the resilience of each network due to inherent uncertainty and lack of data. This work proposes an approach to modeling the resilience of interdependent water delivery systems and the power grid under uncertainty. The interdependency between the two networks is estimated probabilistically based on spatial proximity using the Euclidean distance. The proposed approach is illustrated with a case study of real-world interdependent water and power networks in Shelby County, Tennessee. The resilience curves of the power grid, the water delivery network, and the combined interdependent networks under different recovery schemes are compared. Accounting for the uncertain and dynamic interdependent links can help reduce recovery time by 11% to 60%.