The long-term goal of this research is to provide theoretically justified control strategies to operate autonomous vehicles in spatiotemporal flowfields. The specific objective of this dissertation is to use estimation and nonlinear control techniques to generate decentralized control algorithms that enable motion coordination for multiple autonomous vehicles while operating in a time-varying flowfield. A cooperating team of vehicles can benefit from sharing data and tasking responsibilities. Many existing control algorithms promote collaboration of autonomous vehicles. However, these algorithms often fail to account for the degradation of control performance caused by flowfields. This dissertation presents decentralized multivehicle coordination algorithms designed for operation in a spatially or temporally varying flowfield. Each vehicle is represented using a Newtonian particle traveling in a plane at constant speed relative to the flow and subject to a steering control. Initially, we assume the flowfield is known and describe algorithms that stabilize a circular formation in a time-varying spatially nonuniform flow of moderate intensity. These algorithms are extended by relaxing the assumption that the flow is known: the vehicles dynamically estimate the flow and use that estimate in the control. We propose a distributed estimation and control algorithm comprising a consensus filter to share information gleaned from noisy position measurements, and an information filter to reconstruct a spatially varying flowfield. The theoretical results are illustrated with numerical simulations of circular formation control and validated in outdoor unmanned aerial vehicle (UAV) flight tests.