Quantitative effects of substrate, vegetation, and hydraulic loading on pollutant removal in constructed wetlands

Pollutant removal efficiency in constructed wetlands (CWs) is governed by substrate composition, vegetation, and hydraulic loading rate (HLR). This study conducted a cross-system quantitative analysis of these factors, revealing that laboratory-based CWs studies concentrated on nitrogen and phosphorus nutrients (92.35 %) and oxygen-consuming organic matter (76.47 %). Core indicators, including total nitrogen (TN), total phosphorus (TP), chemical oxygen demand (COD), and ammonium nitrogen (NH 4 + -N), exhibited a median removal efficiency of 59–67 %, with potential negative values observed for TP and NH 4 + -N at low influent concentrations. Natural materials (41.38 %) constituted the highest proportion of substrates, with layered substrate structure (65.06 %) emerging as the predominant approach. Among common single-substrate configurations, gravel showed optimal COD removal (68.67 %), zeolite enhanced nitrogen removal (NH 4 + -N is 40 %), and ceramic improved COD degradation (42.86 %). Notably, composite substrates substantially augment removal efficiency compared to single substrates, particularly in scenarios characterized by low initial removal performance. Plant cultivation increased removal efficiency for TN, TP, COD, and NH 4 + -N by 14.20 %, 16.00 %, 11.55 %, and 17.33 %, respectively, with the most robust conclusion drawn for nitrogen removal enhancement. The optimal HLR for nitrogen and phosphorus removal was 0.5 m 3 /(m 2 ·d), while COD required a lower HLR (0.25 m 3 /(m 2 ·d)) to ensure adequate dissolved oxygen availability. High nitrogen and low carbon influent conditions facilitated HLR reduction, thereby improving overall system efficiency. By quantitatively assessing the impact of substrate selection, vegetation configuration, and HLR on CWs operational performance, we have identified the typical values from laboratory results, providing parameter reference values for engineering applications, and ultimately offering a decision-making basis for transitioning CWs from empirical design to data-driven approaches.