From Literature\nto Action: Analyzing How Nitrogen\nand Phosphorus Removal Responds to Different Design Factors in Free\nWater Surface Constructed Wetlands

Despite the increasing number of studies on nitrogen\n(N) and phosphorus\n(P) removal in free water surface (FWS) wetland systems, there is\nstill a gap in understanding the influence of design variables on\nthe system performance. To address this, we conducted a global meta-analysis\nof 73 studies employing advanced statistical techniques, kinetic models,\nand machine learning along with variable importance analysis. The\nresults indicated that random forest (<i>R</i><sup>2</sup> = 0.55–0.77) and artificial neural network (<i>R</i><sup>2</sup> = 0.5–0.85) were the best fitting models for\nTN and TP removal in pilot-scale and large-scale systems. Moreover,\npermutation importance results using different wetland design variables\nindicated that the inflow concentration, plant coverage, hydraulic\nloading rate, and system area are considered the most important variables\nfor N and P removal under large-scale conditions, while the hydraulic\nretention time, inflow concentration, and water depth are deemed the\nmost important variables under pilot-scale conditions. Furthermore,\nthe removal of N and P was higher in pilot-scale (54.6% and 56.7%)\nsystems compared to that in large-scale (29.0% and 41.9%) systems.\nAlso, the interactions between design variables and the removal process\nof N and P were investigated to better understand the specific roles\nof these variables in improving the removal performance.