Using the Pearl River Estuary (PRE) as an example, the exposure time (θ‾) was calculated to investigate the dominant processes of three-dimensional water exchanges in the estuarine-shelf system and explore the underlying transport pathway. The θ‾, which quantifies the time of water stayed in a defined region, shows clear seasonality under the joint controls from river discharge, tides and wind forcing. It has the lowest domain-average value (∼6 days) in the early summer, then along with weakening of river discharge, it increases to the peak (∼13 days) at the beginning of the fall and remains relatively high during following fall and winter. In contrast, the vertical difference of θ‾ in the water column shows the opposite temporal variability. Spatially, θ‾ increased from the lower to the upper estuary, and in response to the gravitational circulation of surface seaward transport and the bottom intrusion, a larger θ‾ appeared in the bottom layer of water column. Numerical experiments show river discharge predominantly determined the θ‾ compared to the modulation from tides and wind. Based on the water exchange timescale in different subdomains, the transport pathway inside the estuary is controlled by river discharge and tides, which impose a net effect on weakening onshore transport. Outsides the estuary, the wind-driven shelf current modulates the remote influences from the regions far away from the PRE. The spatial structure and temporal variabilities of water exchange identified in this study help to utilize the water exchange capability and realize substance renewal in estuary-shelf regions.