Design framework for sizing free water surface wetlands subject to intermittent wastewater loading in aquaculture facilities

The rapid growth of aquaculture has provided economic growth and food security especially in the developing world, but the pervasive lack of aquaculture pond wastewater treatment before discharge increases coastal water pollution. Although free water surface (FWS) treatment wetlands have potential to treat aquaculture wastewater, the unique harvesting methods used in aquaculture can cause removal variability leading to non-compliance if designed improperly. During harvest, pond wastewater flow and pollutant loading are transient and intermittent. Existing treatment wetland sizing guidelines assume steady-state pollutant transport, steady-flow conditions which are not compatible with aquaculture operation. Three improved design models of increasing complexity are developed for treatment wetlands that simulate incomplete mixing, intermittent loading, and transient wastewater storage. Incomplete mixing is achieved with a tanks-in-series model using three tanks and is capable of simulating variable volumes with different removal constants. The wetland outlet pollutant concentration due to pond releases at different times is captured using the superposition principle. Treatment wetland storage is assumed constant in Model I with transient pollutant transport while the other models assume transient flow and pollutant transport subject to outflow controlled storage (Model II) and inflow/outflow controlled storage (Model III). The models were then linked to three multi-objective, mixed-integer nonlinear optimization models to estimate wetland area. The optimization models are solved using genetic algorithm methodology and the linked simulation-optimization models are referred to as decision support tools (DST) in this study. First-order-second-moment (FOSM) uncertainty analysis was integrated with the combined simulation-optimization model to develop a reliability-based design protocol for treatment wetland design. This approach explicitly considers parametric uncertainty as part of the design process and can be used to address issues related to the treatment variability of wetlands. The FOSM results show that the assumption of steady flow in DST I results in the most conservative design. In contrast, by including variable flows, DST II and DST III have lower design area uncertainty but require more input parameter data which may not be easily available. A decision support flow chart has been developed to provide planners with a methodology to aid in choosing and applying the appropriate DST for their specific conditions.