What drove the nonlinear hypoxia response to nutrient loading in Chesapeake Bay during the 20th century?

Previous data analysis showed that the large expansion of hypoxia in Chesapeake Bay between 1950s and 1980s was correlated to the increased riverine nutrient loading, but the physical and biogeochemical processes driving this hypoxia response need to be better understood. Using a validated coupled hydrodynamic-biogeochemical model, we conducted a hindcast simulation of dissolved oxygen during the 40-year period (1950-1989) when the nutrient loading doubled. The model reproduced the observed decline in O-2 concentration at monitoring stations and the expansion of the hypoxic volume. The peak summer hypoxic volume expanded from similar to 5km(3) during 1950-1969 to similar to 10 km(3) during 1970-1989. To discern how different physical and biochemical processes regulated dissolved O-2, we examined O-2 budget in a fixed control volume of the bottom water most susceptible to hypoxia. The increased water column respiration was found to be the dominant driver of the hypoxia expansion. Further analysis showed a nonlinear response to the nutrient loading. The accumulative hypoxia volume days per unit of nitrate load showed an abrupt (similar to 50 %)jump around 1968. The summer mean hypoxic volume increased with the winter-spring nutrient load, but it was 1.3 km(3)(about 30 %) higher in 1968-1989 than in 1950-1967 at the same nutrient load. This upward shift in hypoxia was caused by the upward shift in the relationship between the water column respiration and winter-spring nutrien tload. Hypoxia suppressed nitrification and denitrification processes in the sediment, amplifying nutrient recycling by 15 % and water column respiration by 12 %. Our modeling analysis demonstrated a feedback mechanism for driving the nonlinear hypoxia response to nutrient loading.