What drives Oregon shelf summer hypoxia

Using coupled biological-physical model based on NPZD-type biological model and 3D coastal ocean model (ROMS) we studied dissolved oxygen (DO) dynamics and hypoxia development on Oregon shelf during April-August of 2002, 2006, and 2008. We found that shelf hypoxia existed during summer months of all three years. It was characterized by variable severity, horizontal and vertical extent, duration and timing, and it was more pronounced in 2002 and 2006. By the means of numerical sensitivity analysis we found out that: inadequate initial DO and NO3 conditions in late-spring 2002 prevented or delayed hypoxia development; offshore and especially northern DO and NO3 boundary conditions are important to simulating hypoxia on the Oregon shelf, this was especially critical for early bottom hypoxia on the shelf north of 45N in 2006; hypoxia occurred earlier in the north in 2006 and in the south (Heceta Bank) in 2002, perhaps, due to different northern boundary conditions for these years; the DO and NO3 conditions at western open boundary located some 400 km offshore are unimportant for DO dynamics in spring-summer. Although DO production due to biological processes is large, physical processes, mostly horizontal advection and diffusion, are responsible for net DO reduction in spring-summer and hypoxia onset in summer on the Oregon shelf. The physical mechanism most responsible for Oregon shelf hypoxia is the coastal upwelling. Diffusive fluxes of NO3 and DO are negligible at northern and southern boundaries of the Oregon shelf and appreciable at the western boundary. In 2006, about two thirds of total April-August DO loss happened in April-May as a result of strong and long-lasting upwelling event.