Modeling and Detection of Structural Change in the Dynamics of DO in a Southeastern Piedmont Impoundment

Proceedings of the 2003 Georgia Water Resources Conference, held April 23-24, 2003, at the University of Georgia.A modified Recursive Prediction Error (RPE) algorithm is used to discover structural change in modeling dynamics of dissolved oxygen (DO) in an aquaculture pond. The small aquaculture pond with a surface area of 1.75 acres, located at Whitehall Forest near the University of Georgia campus, was fertilized on three occasions during the summer of 2000. System responses to nutrient addition have been observed using the University of Georgia’s Environmental Process Control Laboratory (EPCL) for a continuous fivemonth period with 15-minute sampling intervals. A subset of these data – for a period of 36 days from July 1st through August 5th, 2000 – is examined in detail to observe and understand the dynamics of photosynthesis and DO in situ. A mechanistic model performs quite well for describing the concentrations of algae and DO during an algal bloom condition but poorly after the bloom ends. The results of recursive estimation indicate that another process, probably photosynthesis and respiration of macrophytes, may dominate the diurnal variation of DO when the algal biomass in the water column becomes negligible