| |
Treatment wetlands remove contaminants from surface runoff to sustain or improve downstream water uses. In southern Florida USA, the Everglades Stormwater Treatment Areas (STAs) remove phosphorus from agricultural drainage before it is discharged to the ecologically degraded Everglades Protection Area. However, total phosphorus concentrations in STA outflows (cOUT) consistently exceed the regulatory limit (10 ppb), and the drivers of this noncompliance are not clearly understood. Moreover, the current management model (DMSTA), and alternative models proposed in the literature, do not skillfully simulate cOUT dynamics to support effective intervention: The conventional paradigm, which relies on theory-based linear approximations, fails to adequately explain complex treatment behavior. Using nonlinear time series methods, we empirically diagnose the dynamical structure underlying observed cOUT volatility and identify drivers amenable to management intervention. Results indicate that cOUT volatility at STA-3/4 is predominantly systematic due to low-dimensional nonlinear deterministic dynamics. Further, we identify the wetland stage at the STA-3/4 inlet as a strong driver of noncompliance (stronger than the TP concentration at the inlet), suggesting that management of the hydraulic loading rate should be prioritized. Finally, based on these insights we successfully develop an ensemble of parsimonious nonlinear deterministic models that skillfully reproduce long-term dynamics at STA-3/4. |
