Physical transport processes affect the origins of harmful algal blooms in estuaries.

• Numerical model simulations confirm the locations initiating Cochlodinium polykrikoides blooms in the James River. • A simplified mathematical model using ϕ index highlights the role of transport processes in estuarine HAB initiation. • Critical time required for HAB initiation is inversely proportional to relative growth rate. • HABs tend to initiate in locations with relatively large ratios of residence time to volume. • Multiple HAB originating locations can co-exist in an estuary. The effects of physical transport processes on the initiation of harmful algal blooms (HABs) in estuaries were investigated through both mathematical model analysis and numerical model experiments. This study highlights the influence of the flushing effect due to physical transport processes on the location of bloom initiation, which is comparable to or even more important than local processes. The theoretical analysis suggests that the differential flushing effect at different waterbodies due to complex geometry is one of the dominant factors causing inhomogeneous distribution of algal density during HAB initiation. The ratio of residence time to volume is one of the key variables that determine the differential timing of HAB occurrence in estuary-subestuary systems with multiple interconnected waterbodies. As a result, a HAB tends to be observed first in those locations with relatively long residence time and small waterbodies, such as tributaries or areas with large eddies. Multiple unconnected originating locations can co-exist within an estuary. Two three-dimensional model experiments with realistic forcings were conducted to demonstrate the flushing effect on annual Cochlodinium polykrikoides bloom in the lower James River. The results show that while the environmental conditions that affect local processes, such as salinity and temperature, are important in determining the originating locations of HABs, the differential flushing effect is the dominant factor driving the spatial difference in the density of C. polykrikoides in this region during the bloom initiation. This explains why the occurrence of the first bloom in this region is frequently observed in the Lafayette River, a relatively small waterbody with long residence time. Because of the relatively low growth rate of C. polykrikoides and because of the high water-exchange between the mainstem and tributaries of the James River, initial cyst distribution is suggested to have a relatively small impact on originating locations of the bloom compared to flushing effect and salinity, and the HAB originating locations do not have to be in the waterbody with abundant cysts. [ABSTRACT FROM AUTHOR]