Background: The Penjemuran River in Sukarami District, Palembang, is polluted by leachate from the adjacent Sukawinatan Landfill, degrading a fishery-supporting watercourse; substrate optimisation of constructed wetlands for landfill-impacted tropical rivers remains under-studied. Objective: To optimise the substrate of a laboratory-scale Free Water Surface (FWS) constructed wetland for maximal pollutant removal from the Penjemuran River. Methods: A 60×30×50 cm FWS reactor (20 cm substrate, 30 cm water column) planted with Echinodorus palaefolius (4 plants m⁻²; 24-h HRT) was tested with sand–gravel, zeolite and biochar. BOD, COD, TSS, ammonia and pH were measured by SNI methods against Class II limits (PP 22/2021) and analysed by one-way ANOVA, Tukey HSD, Pearson correlation, PCA and pollution indices (α = 0.05). Results: Substrate effects were highly significant for every parameter (p ≤ 0.001; η² = 0.92–0.99). Sand–gravel best removed BOD (89.4%), zeolite best removed COD (81.9%) and TSS (75.0%), and biochar best removed ammonia (83.2%; effluent 0.099 mg/L). BOD and ammonia removal were inversely correlated (r = −0.93), and PC1–PC2 explained 98.6% of the variance. The Pollution Index reached compliance only when adsorptive media were combined. Conclusion: No single medium met all Class II limits; a stratified biochar–zeolite–sand–gravel field-scale system (5×1.5×0.5 m, 1 m³ day⁻¹, IDR 45,414,703) is recommended for rehabilitating landfill-impacted rivers and safeguarding community fisheries.
Copyrights © 2026